NONDESTRUCTIVE TESTING
Third Edition
HANDBOOK
Volume 1
Leak Testing
Technical Editors
Charles N. Jackson, Jr.
Charles N. Sherlock
Editor
Patrick O. Moore
American Society for Nondestructive Testing
NONDESTRUCTIVE TESTING
Third Edition
HANDBOOK
Volume 1
Leak
Testing
Technical Editors
Charles N. Jackson, Jr.
Charles N. Sherlock
Editor
Patrick O. Moore
®
DED
FOUN
1941
American Society for Nondestructive Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
Copyright © 1998
AMERICAN SOCIETY FOR NONDESTRUCTIVE TESTING, INC.
All rights reserved.
No part of this book may be reproduced, stored in a retrieval system or transmitted, in any form or by any means —
electronic, mechanical, photocopying, recording or otherwise — without the prior written permission of the publisher.
Nothing contained in this book is to be construed as a grant of any right of manufacture, sale or use in connection with
any method, process, apparatus, product or composition, whether or not covered by letters patent or registered
trademark, nor as a defense against liability for the infringement of letters patent or registered trademark.
The American Society for Nondestructive Testing, its employees and the contributors to this volume are not responsible
for the authenticity or accuracy of information herein, and opinions and statements published herein do not necessarily
reflect the opinion of the American Society for Nondestructive Testing or carry its endorsement or recommendation.
The American Society for Nondestructive Testing, its employees, and the contributors to this volume assume no
responsibility for the safety of persons using the information in this book.
Library of Congress Cataloging-in-Publication Data
Leak Testing / technical editors, Charles N. Jackson, Jr., Charles N. Sherlock ;
editor, Patrick O. Moore. -- 3rd ed.
p.
cm. — (Nondestructive testing handbook ; v. 1)
Includes bibliographic references and index.
ISBN-13 978-1-57117-071-2
ISBN-10 1-57117-071-5
1. Leak detectors. 2. Gas leakage. I. Jackson, Charles N. II. Sherlock,
Charles N. III. Moore, Patrick O. IV. American Society for Nondestructive
Testing. V. Series: Nondestructive testing handbook (3rd ed.) ; v. 1.
TA165.L34 1998
98-10437
620.1’127--dc21
CIP
ISBN-13: 978-1-57117-071-2 (print)
ISBN-13: 978-1-57117-038-5 (CD)
ISBN-13: 978-1-57117-289-1 (ebook)
Errata
You can check for errata for this and other ASNT publications at
<https://www.asnt.org/errata>.
First printing 05/98
Second printing with revisions 12/04
Third printing 09/07
Fourth printing 03/11
ebook 07/13
Published by the American Society for Nondestructive Testing
PRINTED IN THE UNITED STATES OF AMERICA
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
In memory of
Charles N. Sherlock
(1932–1997)
Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
iii
President’s Foreword
This book is the first volume of the third
edition of the Nondestructive Testing
Handbook. The existence of books such as
Leak Testing is testimony to the dedication
of the American Society for
Nondestructive Testing (ASNT) to its
missions of providing technical
information and instructional materials
and of promoting nondestructive testing
technology as a profession. The series
documents advances in the various
nondestructive testing methods and
provides reference materials for
nondestructive testing educators and
practitioners in the field. ASNT’s hope is
that the third edition will build on the
successes of the past and surpass them by
providing current information about our
rapidly evolving technology.
Leak Testing was written and reviewed
under the guidance of ASNT’s Handbook
Development Committee. The
collaboration between the volunteers and
staff in the this volume has made
productive use of ASNT’s volunteer
resources. Scores of authors and reviewers
have donated thousands of hours to this
volume. A special note of thanks is
extended to Handbook Development
Director Gary Workman, to Leak Testing
Committee Chair Gary Elder, to Technical
Editors Charles Sherlock and Charles
Jackson, to Handbook Coordinators John
Keve and Stuart Tison and to Handbook
Editor Patrick Moore for their dedicated
efforts and commitment in providing this
significant book.
Hussein M. Sadek
ASNT National President (1997–98)
iv
Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
Foreword
The Aims of a Handbook
The volume you are holding in your hand
is the first in the third edition of the
Nondestructive Testing Handbook. Now, with
the beginning of a new series, is a good
time to reflect on the purposes and nature
of a handbook.
Handbooks exist in many disciplines of
science and technology, and certain
features set them apart from other
reference works. A handbook should
ideally give the basic knowledge necessary
for an understanding of the technology,
including both scientific principles and
means of application.
The typical reader may be assumed to
have completed three years of college
toward a degree in mechanical
engineering or materials science and
hence has the background of an
elementary physics or mechanics course.
Occasionally an engineer may be
frustrated by the difficulty of the
discussion in a handbook. That happens
because the assumptions about the reader
vary according to the subject in any given
section. Computer science requires a
different sort of background from nuclear
physics, for example, and it is not possible
for the handbook to give all the
background knowledge that is ancillary to
nondestructive testing.
A handbook offers a view of its subject
at a certain period in time. Even before it
is published, it starts to get obsolete. The
authors and editors do their best to be
current but the technology will continue
to change even as the book goes to press.
Standards, specifications,
recommended practices and inspection
procedures may be discussed in a
handbook for instructional purposes, but
at a level of generalization that is
illustrative rather than comprehensive.
Standards writing bodies take great pains
to ensure that their documents are
definitive in wording and technical
accuracy. People writing contracts or
procedures should consult real standards
when appropriate.
Those who design qualifying
examinations or study for them draw on
handbooks as a quick and convenient way
of approximating the body of knowledge.
Committees and individuals who write or
anticipate questions are selective in what
they draw from any source. The parts of a
handbook that give scientific background,
for instance, may have little bearing on a
practical examination. Other parts of a
handbook are specific to a certain
industry. Although a handbook does not
pretend to offer a complete treatment of
its subject, its value and convenience are
not to be denied.
The present volume is a worthy
beginning for the third edition. The
editors, technical editors and many
contributors and reviewers worked
together to bring the project to
completion. For their scholarship and
dedication I thank them all.
Gary L. Workman
Handbook Development Director
Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
v
Preface
Unfortunately, too many people still have
the impression that leak testing involves
little more than finding a hole in a flat
tire. The development of the helium mass
spectrometer in the days of the
Manhattan Project during the 1940s was
the initial quantum leap in leak testing.
With miniaturization and technological
advances in electronics and hardware,
leak testing has grown into a technology
of great sophistication.
In 1982, the American Society for
Nondestructive Testing (ASNT) published
Leak Testing, the first volume of the
second edition Nondestructive Testing
Handbook. Since then, 3000 copies of that
book have been sold, providing many
leak testing personnel, both technicians
and managers, with a ready source of
reference information.
In May 1990, to determine the general
location of apparent leakage, the National
Aeronautics and Space Administration
had to develop a combination of remote
hydrogen sensors and a multiple channel
mass spectrometer connected to a
computer for numeric readouts during
liquid hydrogen fueling. This illustrates
the versatility of the mass spectrometer
and also points out the need for more
research and development to improve
leak testing monitoring systems.
It is good to have aspirations about
space travel, but the pressing reality of the
moment is the environmental damage we
continue to inflict on our space home,
Earth. We are rapidly destroying the
environment in which we live through
contamination of the air we breathe, the
water we drink and the soil in which we
grow our food.
One of the problems today is the many
storage tanks and ponds that have been
leaking contaminants (all sorts of
petrochemical and petroleum products)
into the ground for years with no
effective continuous leakage monitoring.
Many of these structures were not
adequately leak tested at the time they
were fabricated and, until recently, were
not closely monitored for leakage that
passed into the ground, contaminating
the soil and water supply.
What does leak testing have to do with
all of this? It is the one nondestructive
testing method that can be used to
determine the total leakage rate (quantity
or mass) of undesirable products escaping
vi
from their containers into the
environment. A combination of pressure
change and mass flow in one form or
another has been used for this purpose for
many decades. A good example is the
integrated leakage rate testing of nuclear
containment systems. The existence of
these containment systems and the tests
that proved their total leakage to be
within acceptable limits helped reduce
the environmental damage from the
incident at Three Mile Island. Without
these safeguards, that incident would
have been an environmental catastrophe
such as occurred at Chernobyl in the
Ukraine.
Many combinations of volume change,
tracer gas testing with detector probes,
liquid displacement, ultrasound etc. are
used to test storage tanks. Needed now are
quantitative test techniques sensitive
enough to detect all fluid leakage and yet
reasonably economical for construction of
tank configurations and products. It is
time for development of better leak
testing systems and procedures for these
structures.
More training, qualification and
certification for leak testing personnel will
be implemented when management
realizes that nondestructive testing can
save money and when codes and
standards include such requirements. The
impetus to make it happen will have to
come from the nondestructive testing
community and organizations like ASNT.
The Technical Editors would like to
thank all the ASNT staff and volunteers —
contributors, reviewers and committee
members — who made this book possible.
Charles N. Jackson, Jr.
Charles N. Sherlock
Technical Editors
Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
Editor’s Preface
The third edition of the Nondestructive
Testing Handbook begins as the second
edition did, with the volume Leak Testing.
This third edition volume is indebted to
the preceding edition’s volume in many
ways. Much of the text is the same,
despite significant additions and
alterations.
Published in 1959 by the American
Society for Nondestructive Testing (ASNT),
the first edition of the Nondestructive
Testing Handbook did not cover leak
testing at all. In 1982, the second edition’s
Leak Testing volume was groundbreaking.
Aside from the Leakage Testing Handbook
(1968), written by J.W. Marr for the
National Aeronautics and Space
Administration, there had been no
comprehensive books on the subject.
Although parts of Leak Testing drew on
Marr’s work, on standards published by
sister societies and on literature provided
by equipment manufacturers, Leak Testing
was a highly original contribution to
technical literature. For this reason, the
second edition Leak Testing contained very
few references to other publications.
The technical content of this third
edition volume differs in several ways
from that of the second. (1) New
technology is represented, including
infrared thermography and counterflow
mass spectrometry. (2) Pages have been
added to cover new applications, such as
the inspection of storage tanks. (3) The
text reflects the fact that, for reasons of
environment, fluorocarbon tracer gases
have been regulated. (4) A comprehensive
glossary is provided. (5) An extensive
bibliography lists leak testing
publications, more than some leak testing
practitioners might have expected.
The greatest setback during the
preparation of this volume was the death
in February 1997 of Technical Editor
Charles Sherlock. He contributed many
pages to this volume and edited the first
half through the galley stage. His good
humor and willingness to give freely of
his time and knowledge endeared him to
many ASNT members. The technical
community will continue to miss him for
many years.
After his passing, the task of editing for
technical accuracy was undertaken by
Charles Jackson. ASNT is very fortunate
that he was willing to devote his technical
expertise to this project.
ASNT is likewise indebted to Handbook
Coordinators Stuart Tison and John Keve
and to the technical experts listed at the
end of this foreword. (Please note that
people listed as contributors were also
reviewers but are listed only once, as
contributors.)
It is difficult to overstate the
contributions of staff members Hollis
Humphries-Black and Joy Grimm to the
art, layout and text of the book. I would
also like to thank Publications Manager
Paul McIntire for his support during
design and production.
Patrick O. Moore
Editor
Acknowledgments
Handbook Development
Committee
Gary L. Workman, University of Alabama
in Huntsville
Michael W. Allgaier, GPU Nuclear
Robert A. Baker
Albert S. Birks, AKZO Nobel Chemicals
Richard H. Bossi, Boeing Aerospace
Company
Lawrence E. Bryant, Jr., Los Alamos
National Laboratory
John Stephen Cargill, Pratt & Whitney
William C. Chedister, Circle Chemical
Company
James L. Doyle, Lotis Technologies
Corporation
Matthew J. Golis
Allen T. Green, Acoustic Technology
Group
Robert E. Green, Jr., Johns Hopkins
University
Grover Hardy, Wright-Patterson Air Force
Base
Frank A. Iddings
Charles N. Jackson, Jr.
John K. Keve, DynCorp Tri-Cities Services
Lloyd P. Lemle, Jr.
Xavier P.V. Maldague, University Laval
Paul McIntire, ASNT
Michael L. Mester, Timken Company
Scott D. Miller, Aptech Engineering
Services
Ronnie K. Miller, Physical Acoustics
Corporation
Patrick O. Moore, ASNT
Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
vii
Stanley Ness
Ronald T. Nisbet
Philip A. Oikle, Yankee Atomic Electric
Company
Emmanuel P. Papadakis, Quality Systems
Concepts
Stanislav I. Rokhlin, Ohio State University
J. Thomas Schmidt, J. Thomas Schmidt
Associates
Amos Sherwin, Sherwin, Incorporated
Kermit Skeie, Kermit Skeie Associates
Roderic K. Stanley, Quality Tubing
Philip J. Stolarski, California Department
of Transportation
Holger H. Streckert , General Atomics
Stuart A. Tison, National Institute of
Standards and Technology, Vacuum
Group
Noel A. Tracy, Universal Technology
Corporation
Mark F.A. Warchol, Aluminum Company
of America
George C. Wheeler
Robert Windsor, ASNT
Contributors
Gerald L. Anderson, American Gas and
Chemical Company
John F. Beech, GeoSyntec Consultants
Mark D. Boeckmann, Vacuum Technology,
Incorporated
Betty J.R. Chavez, UE Systems
Phillip T. Cole, Physical Acoustics Limited,
Cambridge
Glenn T. Darilek, Leak Location Services
Gary R. Elder, Gary Elder and Associates
James P. Glover, Graftel
Mark A. Goodman, UE Systems
Charles N. Jackson, Jr.
John K. Keve, DynCorp Tri-Cities Services
Daren L. Laine, Leak Location Services
Leonard F. Laskowski, Solutia,
Incorporated
Robert W. Loveless
Ronnie K. Miller, Physical Acoustics
Corporation
George R. Neff, Isovac Engineering
Jimmie K. Neff, Isovac Engineering
Thomas G. McRae, Laser Imaging Systems
Joseph S. Nitkiewicz, Westinghouse
Electric Corporation
Donald J. Quirk, Fisher Controls
International
Paul B. Shaw, Chicago Bridge and Iron
Company
Charles N. Sherlock
Holger H. Streckert, General Atomics
Philip G. Thayer, Physical Acoustics
Corporation
Stuart A. Tison, National Institute of
Standards and Technology
Carl A. Waterstrat, Varian Vacuum
Products
Gary J. Weil, EnTech Engineering
viii
Reviewers
Michael Bonapfl, University of California
at Lawrence Livermore National
Laboratory
William Baker, Teledyne Hastings
Instruments
John S. Buck, Micro Engineering
Martin Conway, Volumetrics,
Incorporated
Jeffrey F. Cook, Sr., JFC NDE Engineering
Mary Beth DiEleonora, Emerson Electric
Company
Jerry Fruit, Mensor Corporation
Joseph Glatz, Qual-X, Incorporated
Allen T. Green, Acoustic Technology
Group
Tony Heinz, Leak Testing Specialists
Stanislav I. Jakuba, SI Jakub Associates
Edsel O. Jurva, Jurva Leak Testing
David Kailer, NDT International
Robert Koerner, Geosynthetic Research
Institute
Betty Ann Kram, Leybold Inficon
David S. Kupperman, Argonne National
Laboratory
Lloyd P. Lemle, Jr.
Keith Lacy, Westinghouse Electric
Corporation
Arthur F. Mahon, Qual-X, Incorporated
Gregory Markel, Helium Leak Testing,
Incorporated
Michael E. McDaniel, EG&G Florida
Michael Murray, Parker Seals Company
Willis C. Parshall, Jr., FES Division of
Thermo Power Corporation
Paul Pedigo, Inframetrics,
Adrian A. Pollock, Physical Acoustics
Corporation
Allen D. Reynolds
John D. Rhea, Yokogawa Corporation of
America
Tito Y. Sasaki, Quantum Mechanics
Corporation
Todd Sellmer, Westinghouse Engineered
Products
Gary Schaefer, Wallace & Tiernan,
Incorporated
Rod L. Shulver, Realistic Systems Tech
Incorporated
John Snell, Snell & Associates
John Tkach, Cryogenics Technology
Incorporated
John Tyson II, Laser Technology
Incorporated
David R. Vincett, Varian Vacuum Products
William C. Worthington, Leybold Inficon
Fred Wiesinger, Uson L.P.
Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
Contents
Chapter 1. Introduction to Leak
Testing . . . . . . . . . . . . . . . . . . . . 1
Part 1. Nondestructive Testing . . . . 2
Part 2. Management and
Applications of Leak
Testing . . . . . . . . . . . . . . . . 7
Part 3. History of Leak Testing . . . 22
Part 4. Units of Measure for
Nondestructive Testing . . 26
Chapter 2. Tracer Gases in Leak
Testing . . . . . . . . . . . . . . . . . . . 33
Part 1. Introduction to Properties
of Tracer Gases for Leak
Testing . . . . . . . . . . . . . . . 34
Part 2. Mechanisms of Gaseous
Flow through Leaks . . . . . 45
Part 3. Practical Measurement of
Leakage Rates with Tracer
Gases . . . . . . . . . . . . . . . . 48
Part 4. Mathematical Theory of Gas
Flow through Leaks . . . . . 59
Chapter 3. Calibrated Reference
Leaks . . . . . . . . . . . . . . . . . . . . .
Part 1. Calibrated Reference Leaks .
Part 2. Operation of Standard
(Calibrated) Halogen
Leaks . . . . . . . . . . . . . . . .
Part 3. Operation of Standard
(Calibrated) Helium Leaks
Part 4. Calibration of Standard
Reference Leaks . . . . . . . .
71
72
81
86
94
Chapter 4. Safety Aspects of Leak
Testing . . . . . . . . . . . . . . . . . . . 101
Part 1. General Safety Procedures
for Test Personnel . . . . . 102
Part 2. Control of Hazards from
Airborne Toxic Liquids,
Vapors and Particles . . . . 104
Part 3. Flammable Liquids and
Vapors . . . . . . . . . . . . . . 113
Part 4. Electrical and Lighting
Hazards . . . . . . . . . . . . . 116
Part 5. Safety Precautions with Leak
Testing Tracer Gases . . . . 123
Part 6. Safety Precautions with
Compressed Gas
Cylinders . . . . . . . . . . . . 130
Part 7. Safety Precautions in
Pressure and Vacuum
Leak Testing . . . . . . . . . . 133
Part 8. Preparation of Pressurized
Systems for Safe Leak
Testing . . . . . . . . . . . . . . 140
Part 9. Exposure to Toxic
Substances . . . . . . . . . . . 150
Chapter 5. Pressure Change and Flow
Rate Techniques for Determining
Leakage Rates . . . . . . . . . . . . . 153
Part 1. Introduction to Pressure
Instrumentation,
Measurements and
Analysis . . . . . . . . . . . . . 154
Part 2. Pressure Change Leakage
Rate Tests in Pressurized
Systems . . . . . . . . . . . . . 184
Part 3. Pressure Change Tests for
Measuring Leakage in
Evacuated Systems . . . . . 192
Part 4. Flow Rate Tests for
Measuring Leakage Rates
in Systems near
Atmospheric Pressure . . . 205
Chapter 6. Leak Testing of Vacuum
Systems . . . . . . . . . . . . . . . . . . 215
Part 1. The Nature of Vacuum . . . 216
Part 2. Principles of Operation of
Vacuum Systems and
Components . . . . . . . . . 223
Part 3. Materials for Vacuum
Systems . . . . . . . . . . . . . 235
Part 4. Vacuum System
Maintenance and
Troubleshooting . . . . . . .238
Part 5. Equipment and Techniques
for Measuring Pressure in
Vacuum Systems . . . . . . 243
Part 6. Techniques for Detection of
Large Leaks in Operating
Vacuum Systems . . . . . . 254
Part 7. Leak Testing of Vacuum
Systems by Vacuum Gage
Response Technique . . . 261
Part 8. Leak Testing of Systems by
Thermal Conductivity
Techniques . . . . . . . . . . 264
Part 9. Leak Testing of Vacuum
Systems by Ionization
Gage or Pump
Techniques . . . . . . . . . . 267
Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
ix
Chapter 7. Bubble Testing . . . . . . . . 275
Part 1. Introduction to Bubble
Emission Techniques of
Leak Testing . . . . . . . . . . 276
Part 2. Theory of Bubble Testing by
Liquid Immersion
Technique . . . . . . . . . . . 286
Part 3. Bubble Testing by Liquid
Film Application
Technique . . . . . . . . . . . 298
Part 4. Bubble Testing by Vacuum
Box Technique . . . . . . . . 306
Part 5. Procedures and
Applications of Bubble
Testing in Industry . . . . 312
Chapter 8. Techniques and
Applications of Helium Mass
Spectrometry . . . . . . . . . . . . . . 319
Part 1. Principles of Mass
Spectrometer Leak Testing
with Helium Tracer Gas . 320
Part 2. Tracer Probe Technique for
Leak Testing of Evacuated
Objects . . . . . . . . . . . . . 330
Part 3. Hood Technique for Leak
Testing of Evacuated
Objects . . . . . . . . . . . . . 336
Part 4. Accumulation Technique
for Leak Testing of
Evacuated Objects . . . . . 343
Part 5. Detector Probe Technique for
Leak Testing of Pressurized
Objects . . . . . . . . . . . . . 345
Part 6. Bell Jar Technique for Leak
Testing of Pressurized
Objects . . . . . . . . . . . . . 357
Part 7. Accumulation Technique
for Leak Testing of
Pressurized Objects . . . . 360
Chapter 9. Mass Spectrometer
Instrumentation for Leak
Testing . . . . . . . . . . . . . . . . . . . 369
Part 1. Principles of Detection of
Helium Gas by Mass
Spectrometers . . . . . . . . 370
Part 2. Sensitivity and Resolution
of Mass Spectrometer
Helium Leak Detectors . . 385
Part 3. Operation and Maintenance
of Mass Spectrometer
Vacuum System . . . . . . . 392
Chapter 10. Leak Testing with Halogen
Tracer Gases . . . . . . . . . . . . . . 405
Part 1. Introduction to Halogen
Tracer Gases and Leak
Detectors . . . . . . . . . . . . 406
Part 2. Introduction to Techniques
of Halogen Leak Testing . 420
x
Leak Testing
Part 3. Recommended Techniques
for Pressure Leak Testing
with Halogen Detector
Probe . . . . . . . . . . . . . . . 432
Part 4. Industrial Applications of
Halogen Leak Detection . 442
Part 5. Writing Specifications for
Halogen Leak Testing . . . 450
Chapter 11. Acoustic Leak Testing . . 457
Part 1. Principles of Sonic and
Ultrasonic Leak Testing . 458
Part 2. Instrumentation for
Ultrasound Leak Testing 467
Part 3. Ultrasound Leak Testing of
Pressurized Industrial and
Transportation Systems . 474
Part 4. Ultrasound Leak Testing of
Evacuated Systems . . . . . 487
Part 5. Ultrasound Leak Testing of
Engines, Valves, Hydraulic
Systems, Machinery and
Vehicles . . . . . . . . . . . . . 489
Part 6. Electrical Inspection . . . . . 491
Part 7. Ultrasound Leak Testing of
Pressurized Telephone
Cables . . . . . . . . . . . . . . 494
Part 8. Acoustic Emission
Monitoring of Leakage
from Vessels, Tanks and
Pipelines . . . . . . . . . . . . 496
Chapter 12. Infrared Thermographic
Leak Testing . . . . . . . . . . . . . . 505
Part 1. Advantages and Techniques
of Infrared Thermographic
Leak Testing . . . . . . . . . . 506
Part 2. Infrared Leak Testing Using
Emission Pattern
Techniques . . . . . . . . . . 507
Part 3. Leak Testing Using Infrared
Absorption . . . . . . . . . . . 515
Part 4. Infrared Thermographic
Leak Testing Using
Acoustic Excitation . . . . 518
Chapter 13. Leak Testing of
Petrochemical Storage Tanks . . 521
Part 1. Leak Testing of Underground
Storage Tanks . . . . . . . . . 522
Part 2. Leak Testing of Aboveground
Storage Tanks . . . . . . . . . 532
Part 3. Determining Leakage Rate
in Petrochemical
Structures . . . . . . . . . . . . 540
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
Chapter 14. Leak Testing of Hermetic
Seals . . . . . . . . . . . . . . . . . . . . 549
Part 1. Characteristics of Gasketed
Mechanical Hermetic
Seals . . . . . . . . . . . . . . . . 550
Part 2. Characteristics of
Hermetically Sealed
Packages . . . . . . . . . . . . 554
Part 3. Techniques for Gross Leak
Testing of Hermetically
Sealed Devices . . . . . . . . 558
Part 4. Fine Leak Testing of
Hermetically Sealed
Devices with Krypton-85
Gas . . . . . . . . . . . . . . . . 564
Part 5. Fine Leak Testing of
Hermetically Sealed
Devices with Helium
Gas . . . . . . . . . . . . . . . . 574
Chapter 15. Leak Testing Techniques for
Special Applications . . . . . . . . . 579
Part 1. Techniques with Visible
Indications of Leak
Locations . . . . . . . . . . . . 580
Part 2. Primary Containment
Leakage Rate Testing in
the United States Nuclear
Power Industry . . . . . . . 589
Part 3. Leak Testing of Geosynthetic
Membranes . . . . . . . . . . 592
Part 4. Residual Gas
Analysis . . . . . . . . . . . . . 598
Chapter 16. Leak Testing Glossary . . 603
Chapter 17. Leak Testing
Bibliography . . . . . . . . . . . . . . 615
Index
. . . . . . . . . . . . . . . . . . . . . . . . 627
Figure Sources . . . . . . . . . . . . . . . . . . 637
Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
xi
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
C
1
H A P T E R
Introduction to Leak
Testing
Charles N. Sherlock, Willis, Texas
Holger H. Streckert, General Atomics, San Diego,
California (Part 4)
Carl Waterstrat, Varian Vacuum Products, Lexington,
Massachusetts (Part 2)
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
PART 1. Nondestructive Testing
Nondestructive testing (NDT) has been
defined as comprising those test methods
used to examine or inspect a part or
material or system without impairing its
future usefulness.1 The term is generally
applied to nonmedical investigations of
material integrity.
Strictly speaking, this definition of
nondestructive testing includes
noninvasive medical diagnostics. X-rays,
ultrasound and endoscopes are used by
both medical and industrial
nondestructive testing. Medical
nondestructive testing, however, has come
to be treated by a body of learning so
separate from industrial nondestructive
testing that today most physicians never
use the word nondestructive.
Nondestructive testing is used to
investigate specifically the material
integrity of the test object. A number of
other technologies — for instance, radio
astronomy, voltage and amperage
measurement and rheometry (flow
measurement) — are nondestructive but
are not used specifically to evaluate
material properties. Radar and sonar are
classified as nondestructive testing when
used to inspect dams, for instance, but
not when they are used to chart a river
bottom.
Nondestructive testing asks “Is there
something wrong with this material?”
Various performance and proof tests, in
contrast, ask “Does this component
work?” This is the reason that it is not
considered nondestructive testing when
an inspector checks a circuit by running
electric current through it. Hydrostatic
pressure testing is another form of proof
testing and may destroy the test object.
Another gray area that invites various
interpretations in defining nondestructive
testing is future usefulness. Some material
investigations involve taking a sample of
the inspected part for testing that is
inherently destructive. A noncritical part
of a pressure vessel may be scraped or
shaved to get a sample for electron
microscopy, for example. Although future
usefulness of the vessel is not impaired by
the loss of material, the procedure is
inherently destructive and the shaving
itself — in one sense the true “test object”
— has been removed from service
permanently.
The idea of future usefulness is relevant
to the quality control practice of
sampling. Sampling (that is, the use of
2
Leak Testing
less than 100 percent inspection to draw
inferences about the unsampled lots) is
nondestructive testing if the tested sample
is returned to service. If the steel is tested
to verify the alloy in some bolts that can
then be returned to service, then the test
is nondestructive. In contrast, even if
spectroscopy used in the chemical testing
of many fluids is inherently
nondestructive, the testing is destructive if
the samples are poured down the drain
after testing.
Nondestructive testing is not confined
to crack detection. Other discontinuities
include porosity, wall thinning from
corrosion and many sorts of disbonds.
Nondestructive material characterization
is a growing field concerned with material
properties including material
identification and microstructural
characteristics — such as resin curing, case
hardening and stress — that have a direct
influence on the service life of the test
object.
Nondestructive testing has also been
defined by listing or classifying the
various methods.1-3 This approach is
practical in that it typically highlights
methods in use by industry.
Purposes of
Nondestructive Testing
Since the 1920s, the art of testing without
destroying the test object has developed
from a laboratory curiosity to an
indispensable tool of production. No
longer is visual examination of materials,
parts and complete products the principal
means of determining adequate quality.
Nondestructive tests in great variety are in
worldwide use to detect variations in
structure, minute changes in surface
finish, the presence of cracks or other
physical discontinuities, to measure the
thickness of materials and coatings and to
determine other characteristics of
industrial products. Scientists and
engineers of many countries have
contributed greatly to nondestructive test
development and applications.
The various nondestructive testing
methods are covered in detail in the
literature but it is always wise to consider
objectives before plunging into the details
of a method. What is the use of
nondestructive testing? Why do
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
thousands of industrial concerns buy the
testing equipment, pay the subsequent
operating costs of the testing and even
reshape manufacturing processes to fit the
needs and findings of nondestructive
testing?
Modern nondestructive tests are used
by manufacturers (1) to ensure product
integrity and, in turn, reliability; (2) to
avoid failures, prevent accidents and save
human life; (3) to make a profit for the
user; (4) to ensure customer satisfaction
and maintain the manufacturer’s
reputation; (5) to aid in better product
design; (6) to control manufacturing
processes; (7) to lower manufacturing
costs; (8) to maintain uniform quality
level; and (9) to ensure operational
readiness.
These reasons for widespread profitable
use of nondestructive testing are sufficient
in themselves, but parallel developments
have contributed to its growth and
acceptance.
Increased Demand on Machines
In the interest of greater speed and rising
costs of materials, the design engineer is
always under pressure to reduce weight.
This can sometimes be done by
substituting aluminum or magnesium
alloys for steel or iron, but such light
alloy parts are not of the same size or
design as those they replace. The
tendency is also to reduce the size. These
pressures on the designer have subjected
parts of all sorts to increased stress levels.
Even such commonplace objects as
sewing machines, sauce pans and luggage
are also lighter and more heavily loaded
than ever before. The stress to be
supported is seldom static. It often
fluctuates and reverses at low or high
frequencies. Frequency of stress reversals
increases with the speeds of modern
machines and thus parts tend to fatigue
and fail more rapidly.
Another cause of increased stress on
modern products is a reduction in the
safety factor. An engineer designs with
certain known loads in mind. On the
supposition that materials and
workmanship are never perfect, a safety
factor of 2, 3, 5 or 10 is applied. Because
of other considerations though, a lower
factor is often used, depending on the
importance of lighter weight or reduced
cost or risk to consumer.
New demands on machinery have also
stimulated the development and use of
new materials whose operating
characteristics and performance are not
completely known. These new materials
create greater and potentially dangerous
problems. As an example, there is a record
of an aircraft’s being built from an alloy
whose work hardening, notch resistance
and fatigue life were not well known.
After relatively short periods of service
some of these aircraft suffered disastrous
failures. Sufficient and proper
nondestructive tests could have saved
many lives.
As technology improves and as service
requirements increase, machines are
subjected to greater variations and to
wider extremes of all kinds of stress,
creating an increasing demand for
stronger materials.
Engineering Demands for Sounder
Materials
Another justification for the use of
nondestructive tests is the designer’s
demand for sounder materials. As size and
weight decrease and the factor of safety is
lowered, more and more emphasis is
placed on better raw material control and
higher quality of materials, manufacturing
processes and workmanship.
An interesting fact is that a producer of
raw material or of a finished product
frequently does not improve quality or
performance until that improvement is
demanded by the customer. The pressure
of the customer is transferred to
implementation of improved design or
manufacturing. Nondestructive testing is
frequently called on to deliver this new
quality level.
Public Demands for Greater Safety
The demands and expectations of the
public for greater safety are apparent
everywhere. Review the record of the
courts in granting higher and higher
awards to injured persons. Consider the
outcry for greater automobile safety, as
evidenced by the required use of auto
safety belts and the demand for air bags,
blowout proof tires and antilock braking
systems. The publicly supported activities
of the National Safety Council,
Underwriters Laboratories, the
Environmental Protection Agency and the
Federal Aviation Administration in the
United States, and the work of similar
agencies abroad, are only a few of the
ways in which this demand for safety is
expressed. It has been expressed directly
by the many passengers who cancel
reservations immediately following a
serious aircraft accident. This demand for
personal safety has been another strong
force in the development of
nondestructive tests.
Rising Costs of Failure
Aside from awards to the injured or to
estates of the deceased and aside from
costs to the public (e.g. evacuation due to
chemical leaks), consider briefly other
Introduction to Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
3
factors in the rising costs of mechanical
failure. These costs are increasing for
many reasons. Some important ones are
(1) greater costs of materials and labor;
(2) greater costs of complex parts;
(3) greater costs due to the complexity of
assemblies; (4) greater probability that
failure of one part will cause failure of
others due to overloads; (5) trend to lower
factors of safety; (6) probability that the
failure of one part will damage other parts
of high value; and (7) part failure in an
automatic production machine, shutting
down an entire high speed, integrated,
production line. When production was
carried out on many separate machines,
the broken one could be bypassed until
repaired. Today, one machine is tied into
the production of several others. Loss of
such production is one of the greatest
losses resulting from part failure.
Applications of
Nondestructive Testing
4
can be completely characterized in terms
of five principal factors: (1) energy source
or medium used to probe object (such as
X-rays, ultrasonic waves or thermal
radiation); (2) nature of the signals, image
and/or signature resulting from
interaction with the object (attenuation of
X-rays or reflection of ultrasound, for
example); (3) means of detecting or
sensing resultant signals (photoemulsion,
piezoelectric crystal or inductance coil);
(4) method of indicating and/or recording
signals (meter deflection, oscilloscope
trace or radiograph); and (5) basis for
interpreting the results (direct or indirect
indication, qualitative or quantitative and
pertinent dependencies).
The objective of each method is to
provide information about the following
material parameters:
1. discontinuities and separations (cracks,
voids, inclusions, delaminations etc.);
2. structure or malstructure (crystalline
structure, grain size, segregation,
misalignment etc.);
3. dimensions and metrology (thickness,
diameter, gap size, discontinuity size
etc.);
4. physical and mechanical properties
(reflectivity, conductivity, elastic
modulus, sonic velocity etc.);
5. composition and chemical analysis
(alloy identification, impurities,
elemental distributions etc.);
6. stress and dynamic response (residual
stress, crack growth, wear, vibration
etc.); and
7. signature analysis (image content,
frequency spectrum, field
configuration etc.).
Nondestructive testing is a branch of the
materials sciences that is concerned with
all aspects of the uniformity, quality and
serviceability of materials and structures.
The science of nondestructive testing
incorporates all the technology for
detection and measurement of significant
properties, including discontinuities, in
items ranging from research specimens to
finished hardware and products. By
definition, nondestructive techniques are
the means by which materials and
structures may be inspected without
disruption or impairment of serviceability.
Using nondestructive testing, internal
properties of hidden discontinuities are
revealed or inferred by appropriate
techniques.
Nondestructive testing is becoming an
increasingly vital factor in the effective
conduct of research, development, design
and manufacturing programs. Only with
appropriate use of nondestructive testing
techniques can the benefits of advanced
materials science be fully realized.
However, the information required for
appreciating the broad scope of
nondestructive testing is available in
many publications and reports.
Terms used in this block are defined in
Table 1 with respect to specific objectives
and specific attributes to be measured,
detected and defined.
The limitations of a method include
conditions required by that method:
conditions to be met for technique
application (access, physical contact,
preparation etc.) and requirements to
adapt the probe or probe medium to the
object examined. Other factors limit the
detection and/or characterization of
discontinuities, properties and other
attributes and limit interpretation of
signals and/or images generated.
Classification of Methods
Classification Relative to Test
Object
In a report, the National Materials
Advisory Board (NMAB) Ad Hoc
Committee on Nondestructive Evaluation
adopted a system that classified methods
into six major categories: visual,
penetrating radiation, magnetic-electrical,
mechanical vibration, thermal and
chemical-electrochemical.3 Each method
Nondestructive testing methods may be
classified according to how they detect
indications relative to the surface of a test
object. Surface methods include liquid
penetrant testing, visual testing, grid and
moiré testing, holography and
shearography. Surface/near-surface
methods include tap, potential drop,
Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
magnetic particle and electromagnetic
testing. When surface or
surface/near-surface methods are applied
during intermediate manufacturing
processes, they provide preliminary
assurance that volumetric methods
performed on the completed object or
component will reveal few if any
rejectable discontinuities, that is, flaws.
Volumetric methods include radiography,
ultrasonic testing, acoustic emission
testing, certain infrared thermographic
techniques and less familiar methods such
as acoustoultrasonic testing and magnetic
TABLE 1. Objectives of nondestructive testing methods.
Objectives
Attributes Measured or Detected
Discontinuites and separations
Surface anomalies
Surface connected anomalies
Internal anomalies
roughness; scratches; gouges; crazing; pitting; inclusions and imbedded foreign material
cracks; porosity; pinholes; laps; seams; folds; inclusions
cracks; separations; hot tears; cold shuts; shrinkage; voids; lack of fusion; pores; cavities;
delaminations; disbonds; poor bonds; inclusions; segregations
Structure
Microstructure
Matrix structure
Small structural anomalies
Gross structural anomalies
molecular structure; crystalline structure and/or strain; lattice structure; strain; dislocation; vacancy;
deformation
grain structure, size, orientation and phase; sinter and porosity; impregnation; filler and/or
reinforcement distribution; anisotropy; heterogeneity; segregation
leaks (lack of seal or through-holes); poor fit; poor contact; loose parts; loose particles; foreign objects
assembly errors; misalignment; poor spacing or ordering; deformation; malformation; missing parts
Dimensions and metrology
Displacement; position
Dimensional variations
Thickness; density
linear measurement; separation; gap size; discontinuity size, depth, location and orientation
unevenness; nonuniformity; eccentricity; shape and contour; size and mass variations
film, coating, layer, plating, wall and sheet thickness; density or thickness variations
Physical and mechanical properties
Electrical properties
Magnetic properties
Thermal properties
Mechanical properties
Surface properties
resistivity; conductivity; dielectric constant and dissipation factor
polarization; permeability; ferromagnetism; cohesive force
conductivity; thermal time constant and thermoelectric potential
compressive, shear and tensile strength (and moduli); Poisson’s ratio; sonic velocity; hardness; temper and
embrittlement
color; reflectivity; refraction index; emissivity
Chemical composition and analysis
Elemental analysis
Impurity concentrations
Metallurgical content
Physiochemical state
detection; identification, distribution and/or profile
contamination; depletion; doping and diffusants
variation; alloy identification, verification and sorting
moisture content; degree of cure; ion concentrations and corrosion; reaction products
Stress and dynamic response
Stress; strain; fatigue
Mechanical damage
Chemical damage
Other damage
Dynamic performance
heat treatment, annealing and cold work effects; residual stress and strain; fatigue damage and life (residual)
wear; spalling; erosion; friction effects
corrosion; stress corrosion; phase transformation
radiation damage and high frequency voltage breakdown
crack initiation and propagation; plastic deformation; creep; excessive motion; vibration; damping; timing of
events; any anomalous behavior
Signature analysis
Electromagnetic field
Thermal field
Acoustic signature
Radioactive signature
Signal or image analysis
potential; strength; field distribution and pattern
isotherms; heat contours; temperatures; heat flow; temperature distribution; heat leaks; hot spots
noise; vibration characteristics; frequency amplitude; harmonic spectrum and/or analysis; sonic and/or
ultrasonic emissions
distribution and diffusion of isotopes and tracers
image enhancement and quantization; pattern recognition; densitometry; signal classification, separation
and correlation; discontinuity identification, definition (size and shape) and distribution analysis;
discontinuity mapping and display
Introduction to Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
5
resonance imaging. Through-boundary
methods described include leak testing,
some infrared thermographic techniques,
airborne ultrasonic testing and certain
techniques of acoustic emission testing.
Other less easily classified methods are
material identification, vibration analysis
and strain gaging.
No one nondestructive testing method
is all-revealing. That is not to say that one
method or technique of a method cannot
be adequate for a specific object or
component. However, in most cases it
takes a series of test methods to do a
complete nondestructive test of an object
or component. For example, if surface
cracks must be detected and eliminated
and the object or component is made of
ferromagnetic material, then magnetic
particle would be the obvious choice. If
that same material is aluminum or
titanium, then the choice would be liquid
penetrant or electromagnetic testing.
However, for either of these situations, if
internal discontinuities were to be
detected, then ultrasonics or radiography
would be the selection. The exact
technique in either case would depend on
the thickness and nature of the material
and the type or types of discontinuities
that must be detected.
manufacturing processes are within design
performance requirements. It should
never be used in an attempt to obtain
quality in a product by using
nondestructive testing at the end of a
manufacturing process. This approach will
ultimately increase production costs.
When used properly, nondestructive
testing saves money for the manufacturer.
Rather than costing the manufacturer
money, nondestructive testing should add
profits to the manufacturing process.
Value of Nondestructive
Testing
The contribution of nondestructive
testing to profits has been acknowledged
in the medical field and computer and
aerospace industries. However, in
industries such as heavy metals, though
nondestructive testing may be grudgingly
promoted, its contribution to profits may
not be obvious to management.
Nondestructive testing is sometimes
thought of as a cost item only. One
possible reason is industry downsizing.
When a company cuts costs, two
vulnerable areas are quality and safety.
When bidding contract work, companies
add profit margin to all cost items,
including nondestructive testing, so a
profit should be made on the
nondestructive testing. However, when
production is going poorly and it is
anticipated that a job might lose money,
it seems like the first corner that
production personnel will try to cut is
nondestructive testing. This is
accomplished by subtle pressure on
nondestructive testing technicians to
accept a product that does not quite meet
a code or standard requirement. The
attitude toward nondestructive testing is
gradually improving as management
comes to appreciate its value.
Nondestructive testing should be used
as a control mechanism to ensure that
6
Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
PART 2. Management and Applications of Leak
Testing4,5
Functions of Leak Testing
Leak testing is a form of nondestructive
testing used in either pressurized or
evacuated systems and components for
detection and location of leaks and for
measurement of fluid leakage. The word
leak refers to the physical hole that exists
and does not refer to the quantity of fluid
passing through that hole. A leak may be
a crack, crevice, fissure, hole or
passageway that, contrary to what is
intended, admits water, air or other fluids
or lets fluids escape (as with a leak in a
roof, gas pipe or ship). The word leakage
refers to the flow of fluid through a leak
without regard to physical size of the hole
through which flow occurs. Fluid denotes
any liquid or gas that can flow.
Surface nondestructive testing methods
or volumetric nondestructive testing
methods often reveal through-wall leaks
to a nondestructive testing technician.
However, it would not be economical to
perform a complete surface liquid
penetrant test of an object or component
in order to detect existing leaks. Many of
the penetrant indications would not be
leaks through the wall. Applying the
liquid penetrant to one surface and the
developer to the opposite surface would
increase the probability that only leaks
would be detected, but this liquid
penetrant technique is a leak test. This
complete dependency only on capillary
action to reveal leaks still would not
necessarily be proof that all leaks were
revealed. Adding even a small differential
pressure to aide that capillary action
would further enhance this leak testing
technique’s sensitivity.
Surface methods such as magnetic
particle would be of little value in
revealing leaks because they indicate
linear discontinuities such as cracks or
nonfusion, not through-wall leaks.
Volumetric methods such as radiography
or ultrasonic testing might be useful in
revealing the exact location of a
difficult-to-pinpoint leak, but only after
that leak is detected and known to exist.
A volumetric method such as acoustic
emission has leak testing techniques
useful in pinpointing leaks but such
techniques have rather limited test
sensitivity. Infrared thermography is
another method whose techniques are
directly related to leak testing. Other more
specialized nondestructive testing
methods previously mentioned would be
of little use in detecting or pinpointing
leaks. In the environment of high vacuum
technology for things such as computer
chip production, X-ray tubes, linear
accelerators for both high voltage X-rays
and physics research for gravitational
waves and quarks, the main applicable
nondestructive testing method is leak
testing. Thus, leak testing and methods
and techniques of leak testing must be
included as a part of the nondestructive
testing field.
When the specification for the
manufacture of an object or component
has a required minimum leak size that
must be detected and/or has a required
maximum total leakage rate that must be
proven, then a leak testing method or
technique of a leak testing method must
be performed to comply with that
specification requirement. No other
nondestructive testing method could be
substituted to fulfill that requirement.
Reasons for Leak Testing
Leaks are special types of anomalies that
can have tremendous importance where
they influence the safety or performance
of engineered systems. The operational
reliability of many devices is greatly
reduced if enough leakage exists. Leak
testing is performed for three basic
reasons: (1) to prevent material leakage
loss that interferes with system operation;
(2) to prevent fire, explosion and
environmental contamination hazards or
nuisances caused by accidental leakage;
and (3) to detect unreliable components
and those whose leakage rates exceed
acceptance standards.
The purposes of leak testing are to
ensure reliability and serviceability of
components and to prevent premature
failure of systems containing fluids under
pressure or vacuum. Nondestructive
methods for rapid leak testing of
pressurized or evacuated systems and of
sealed components are thus of great
industrial and military importance.
Introduction to Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
7
Relationship of Leak
Testing to Product
Serviceability
Measuring Leakage Rates
to Characterize Individual
Leaks
Most types of nondestructive tests are
designed to aid in evaluating serviceability
of materials, parts and assemblies. Tests
are used for determining integrity of
structure, measuring thickness or
indicating the presence of internal and
surface anomalies. For most
nondestructive test methods evaluation is
indirect; the quantities measured have to
be properly correlated to the serviceability
characteristics of the material in question.
Thus, the use of indirect tests depends on
the interpretation of the test results. Leak
testing procedures, on the other hand,
facilitate direct evaluation. The measured
leakage rate represents the physical effect
of a faulty condition and thus requires no
further analysis for practical assessment.
The flow of fluid through a leak typically
results from a pressure differential or a
concentration differential of a gaseous
constituent that acts across the pressure
boundary. The flow characteristics of a
leak are often described in terms of the
conductance of the leak. The leak
represents a physical hole with some
equivalent length and internal crosssectional area or diameter. However,
because a leak is not manufactured
intentionally into a product or system,
the leak hole dimensions are generally
unknown and cannot be determined by
nondestructive tests. Therefore, in leak
testing, the quantity used to describe the
leak is the measured leakage rate.
The leakage rate depends on the
pressure differential that forces fluid
through the leak passageway. The higher
this pressure difference, the greater the
leakage rate through a given leak.
Therefore, leakage measurements of the
same leak under differing pressure
conditions can result in differing values of
mass flow rate. The leak conductance is
defined both by the leakage rate and the
pressure differential across the leak. Thus,
conductance or leakage rate at a given
pressure for a particular tracer fluid should
always be specified in reporting and
interpreting the results of a leak test.
Determination of Overall
Leakage Rates through
Pressure Boundaries
Many leak tests of large vessels or systems
are concerned with the determination of
the rate at which a liquid, gas or vapor
will penetrate through their pressure
boundaries. Leakage may occur from any
location within a component, assembly or
system to points outside the boundary, or
from external regions to points within a
volume enclosed by a pressure boundary.
When a fluid flows through a small leak,
the leakage flow rate depends on (1) the
geometry of the leak, (2) the nature of the
leaking fluids and (3) the prevailing
conditions of fluid pressure, temperature
and type of flow. For purposes of leak
testing, an easily detectable gas or liquid
tracer fluid may be used, rather than air
or the system operating fluid. Leakage
typically occurs as a result of a pressure
differential between the two regions
separated by the pressure boundary.
The term minimum detectable leakage
refers to the smallest fluid flow rate that
can be detected. The leakage rate is
sometimes referred to as the mass flow
rate. In the case of gas leakage, the
leakage rate describes the number of
molecules leaking per unit of time, if the
gas temperature is constant, regardless of
the nature of the tracer gas used in leak
testing. When the nature of the leaking
gas and the gas temperature are known, it
is possible to use the ideal gas laws to
determine the actual mass of the leakage.
8
Leak Testing
Ensuring System Reliability
through Leak Testing
One important reason for leak testing is to
measure the reliability of the system
under test. Leak testing is not a direct
measure of reliability, but it might show a
fundamental fault of the system by a
higher than expected leakage rate
measurement. A high rate of leakage from
mechanical connections might indicate
that a gasket is improperly aligned or
missing. In the same manner, a high
leakage value might show the presence of
a misaligned or misthreaded flange.
Therefore, it is possible to detect
installation errors by high leakage values.
(However, the absence of high leakage
does not necessarily indicate the absence
of improperly installed components.)
Leakage measurements to detect
installation errors need not be extremely
sensitive, because the leakage rates to be
expected from serious error will be
relatively large (10–1 to 10–5 Pa·m3·s–1 or 1
to 10–4 std cm3·s–1). Thus, leak locations
can usually be detected easily.
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
For practical discussions, a small leak is
often defined as having a low leakage rate,
that is, less than that which ensures water
tightness, about 10–5 Pa·m3·s–1
(10–4 std cm3·s–1). Leaks greater than
10–5 Pa·m3·s–1 (10–4 std cm3·s–1) are
considered large.
Leak Testing to Detect
Material Flaws
Many leaks are caused by material
anomalies such as cracks and fissures.
Some of these can be detected by
measurement of leakage rates. Other leaks
can be detected by discontinuity detection
techniques that identify leak locations.
However, neither of these two leak testing
technique categories will detect all
anomalies. Leak testing is therefore
complementary to other nondestructive
testing methods used to find and evaluate
basic material anomalies.
Because service reliability is not
necessarily a direct function of the leakage
in a system, it is difficult to establish an
acceptance level for leakage rate. The
decision may be influenced by the fact
that increased leak testing sensitivity may
detect only a small number of additional
leaks at considerable added cost. This is
because most leaks in welded, brazed and
mechanical joints tend to be relatively
large. This is partly due to the clogging of
smaller leaks by water vapor and liquids
that occurs in parts exposed to industrial
processes or to the atmosphere. The only
case where very small leaks of less than
10–8 Pa·m3·s–1 (10–7 std cm3·s–1) are
encountered is in parts that receive special
clean room treatment during
manufacture.
Specifying Desired
Degrees of Leak Tightness
In industry, the term leaktight has taken
on a variety of meanings. A water bucket
is tight if it does not allow easily
detectable quantities of water to leak out.
A high vacuum vessel is tight if the rate of
apparent leakage into the system cannot
be indicated with the equipment on
hand. One might even consider that a
gravel truck is leaktight so long as there
are no openings in the truck bed large
enough to allow the smallest nugget to
escape. The degree of leak tightness
depends on the individual situation. Leak
tightness requires that the leakage flow be
too small to be detected. However, leak
tightness is a relative term. Therefore, it
becomes a necessity to establish a
practical level of leak testing sensitivity
for any given component under test.
Thus, nothing is leaktight except by
comparison to a standard or specification.
Even then, the measured degree of leak
tightness can be ensured only at the time
of leak testing and under specific leak
testing conditions. Later operation at
higher pressures or temperatures might
open leaks.
Avoiding Impractical
Specifications for Leak
Tightness
Aiming at absolute tightness is an
academic endeavor. In practice, all that
can be asked for is a more or less stringent
degree of tightness selected according to
the application requirements. Nothing
made by man can truly be considered to
be absolutely leaktight. Even in the
absence of minute porosities, the
permeation of certain gases through
metals, crystals, polymers and glasses still
exists.
Thus, it is necessary to establish a
practical leakage rate that is acceptable for
a given component under test. A
preliminary decision has to be made
concerning the definition of leak
tightness for the particular situation.
Because leak tightness is a relative term
and has no absolute meaning, the
sensitivity of the available leak testing
equipment is a practical guide to
attainable levels of leak testing sensitivity.
Any increase in required sensitivity of leak
testing increases the time required for leak
testing and increases test cost. This
increase in cost of leak testing reaches a
maximum when the leakage specification
is given in such impractical terms as no
detectable leakage, no measureable leakage,
no leakage and zero leakage.
Impractical leak testing specifications
are expensive to implement. They are also
very confusing unless the leak testing
method is precisely described. With
specifications in impractical terms, the
leak testing operator is always working
against background instrument noise. He
must then decide whether the leakage
reading obtained is caused by the random
fluctuations of test instruments or by the
actual detection of specific leakage. It is
much easier to discriminate whether a
measured leakage rate is above or below a
given standard than to discriminate
leakage from random instrument noise. It
is therefore suggested that, when
specified, zero leakage be defined as a
measurable quantitative value of leakage
rate that is insignificant in the operation
of the system. Such a definition allows the
system or the measurement sensitivity to
be compared with a flow through a
standard physical leak. In this way, a
Introduction to Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
9
qualification of the system performance
acceptability can be made during the test
operation.
Specifying Leak Testing
Requirements to Locate
Every Leak
Occasionally it is desirable to locate every
existing leak irrespective of size for the
following reasons.
1. Stress leaks have a habit of growing,
i.e., very small leaks may become very
troublesome later, after repeated
stressing.
2. High temperature leaks may be very
small at test temperature but may
have higher leakage rates at system
operating temperatures.
3. Temperature cycling to either high or
cryogenic levels usually creates stress
that results in change of leakage rates.
The criterion whereby a decision is
made whether or not to seek greater
reliability should be the ratio of cost of
the leak testing procedure to the number
of leaks found. For example, improving
leak testing reliability from 10–6 Pa·m3·s–1
(10–5 std cm3·s–1) to a reliability of
10–7 Pa·m3·s–1 (10–6 std cm3·s–1) may not
be justified. The cost of obtaining the
small increase in reliability may be
prohibitive in relation to the value of the
increase in detection reliability.
The expected leak tightness of sealing
operations that will be used to isolate the
system during leak testing must also be
considered. The leak testing specification
should be written with advice from an
experienced engineer who makes a
judgment of the reasonable value of
allowable leakage rate. Factors to be
considered include the leak testing
method and technique; type, size and
complexity of the system under test; and
the service requirements and operating
conditions under which the tested system
will be used.
Specifying Sensitivity of
Leak Testing for Practical
Applications
In specifying the sensitivity of the leak
testing technique, an optimum leakage
sensitivity value should be sought first.
Large deviations from this optimum value
could increase the cost and the difficulty
of measuring the leakage rate. Secondly,
any increase in the sensitivity specified for
a particular leakage test automatically
increases the cost of leak testing.
Therefore, a compromise has to be
10
Leak Testing
reached between testing cost and leakage
tolerance. Thirdly, the sensitivity required
in leak testing depends on the particular
effects of leakage that must be controlled
or eliminated, as illustrated in the
following examples. Finally, the language
in which the leak testing specification is
written should be easy to interpret and to
implement in testing, to ensure that
management’s goals are achieved by the
leak test.
Specifying Tightness Required to
Control Material Loss by Leakage
The first consideration in specifying the
leak tightness required of a fluid
containment system is to ensure that the
system does not leak sufficient material to
cause system failure during the
operational life of the system. Then the
largest leakage rate is the allowable total
leakage divided by the operational life of
the system. Of course, conversion might
have to be made between numerical
values for the tracer gas leakage during
leak testing and those for the material
leakage under system operation
conditions.
Specifying Tightness Required to
Control Environmental
Contamination by Leakage
Contamination failure of a system might
cause environmental damage, personnel
hazard or degraded appearance. The
environmental damage to a system may
be caused by material leaking either into
or out of the system. For example, system
damage may be caused to a liquid rocket
motor when the oxidizer leaks out of the
storage tank and reacts with parts of the
motor. On the other hand, electronic
components can fail when air or water
vapor enters a hermetically sealed
protective container.
It is sometimes difficult to calculate the
very small amount of material necessary
to cause a contamination failure to occur.
However, in most cases, such calculations
are not impossible if the failure can be
defined. For example, if some decision
can be made as to the allowable amount
of reaction between the oxidizer and the
rocket engine parts, the maximum
acceptable rate of total leakage of oxidizer
from the storage tank can be defined.
Similarly, in an electronic component, if
failure results from adsorption of a
monolayer of leaking molecules on the
surface, then knowing that 1015 molecules
form one monolayer on a square
centimeter of surface makes it possible to
calculate the allowable leakage rate for
this particular component. If failure
results from a pressure rise, then the
maximum allowable pressure, the planned
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
system operation time and system volume
are all that are necessary for calculation of
the allowable leakage rate.
Specifying Tightness Required to
Avoid Personnel Hazard Caused
by Fluid Leakage
Material leakage can cause personnel
hazard during system operation. If the
tolerable concentrations are known, and
these are often reported in literature, it is
again quite easy to calculate the
maximum tolerable equipment leakage
rate.
Specifying Tightness Required to
Avoid Undesirable Appearance
Caused by Leakage
An appearance specification is a
specification for maximum leakage that is
made because leakage of a higher value
will spoil the appearance of the system.
Appearance is often specified when no
more stringent specification is necessary.
A specification for leakage of oil out of
the oil pan of a new car is a good
example. This leakage specification may
not be caused by concern that too much
oil will be lost or that damage to the car
motor will occur; instead, it is specified
because the prospective buyer would not
be inclined to buy a car that is dripping
oil onto the showroom floor.
Specifying Tightness Required to
Ensure Continuing System
Operation
When appearance sets the allowable
leakage of the system, the leakage is often
only a nuisance. However, even leaks that
are largely a nuisance may alter the
effectiveness of the total system. For
example, during the East Coast power
blackout in the United States on
November 9, 1965, a large steam
generator failed during the shutdown
because the auxiliary steam supply used
for lubrication purposes was not available.
This steam supply had been shut off
earlier by workers who were bothered by
excessive leakage of steam through some
valve packing. This steam leakage was not
critical, but it was enough of a nuisance
that the system was shut down for repair.
The repair did not take place in time and
the bearings of the generator burned out
during emergency shutdown of the
system.
Definition of Leak Detector
and Leak Test Sensitivity
A leak detector’s sensitivity is a measure of
the concentration or flow rate of tracer
gas that gives a minimum measureable
leak signal. Sensitivity depends on the
minimum detectable number of tracer gas
molecules entering the detector. The
sensitivity of a leak detector is
independent of the pressure in the system
being tested, provided that time is ignored
as a test factor.
Leak test sensitivity refers to the
minimum detectable amount of leakage
that will occur in a specific period of time
under specified leak test conditions. It is
necessary to state both the leakage rate
and the prevailing test conditions to
properly define leak test sensitivity in
terms of the smallest physical size leak
that can be detected. To avoid confusion,
a set of standard leak test conditions is
required.
Standard Conditions for Leak
Testing
The set of conditions most commonly
accepted as standard for pressure
measurement is that of dry air at 25 °C
(77 °F), for a pressure differential between
one standard atmosphere and a vacuum
(a standard atmosphere is roughly 100 kPa
or precisely 101.325 kPa). For practical
purposes, the vacuum need be no better
than 0.01 of an atmosphere or 1 kPa (0.15
lbf·in.–2). When a leak is being described
and only the leakage rate is given, it is
assumed that the leakage rate refers to
leakage at standard conditions. The
sensitivity of a leak testing instrument is
synonymous with the minimum
detectable leakage or minimum flow rate
the instrument can detect. These minima
are independent of leak testing
conditions. When the instrument is
applied to a test, the leak testing
sensitivity depends on existing conditions
of pressure differential, temperature and
fluid type in addition to the instrument
sensitivity. However, the leak test
instrument should be more sensitive by at
least a factor of 2 than the minimum
leakage to be detected, to ensure
reliability and reproducibility of
measurements.
Example of Sensitivity and
Difficulty of Bubble Leak
Testing
Each modification of a leak testing
procedure has an optimum sensitivity
value at which it is most readily used.
Introduction to Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
11
Deviation from this optimum value of
sensitivity makes it more difficult to
perform the measurement and decreases
confidence in the results. Figure 1 shows
the influence of leak testing sensitivity
level on the ease of operation of test
equipment. In most cases, after reaching a
plateau, further increase of sensitivity
rapidly decreases the ease of operation.
Bubble testing by immersion in water is
an example of how the optimum value
affects the ease of performing the test.
The bubble testing sensitivity range
extends from 10–2 to 10–5 Pa·m3·s–1
(10–1 to 10–4 std cm3·s–1). In measuring for
10–2 Pa·m3·s–1 (10–1 std cm3·s–1) leaks, a
component may be placed in water and
observed quickly. Bubbles may emerge
from the pressurized component at such a
rapid rate that there is no question of the
existence of a leak. When checking for
leaks in the range of 10–3 to 10–4 Pa·m3·s–1
(10–2 to 10–3 std cm3·s–1), the operator
must be sure that the test object or
component is submerged long enough for
any bubbles coming from crevices to have
a chance to collect and rise. When
locating leaks in the 10–5 Pa·m3·s–1
(10–4 std cm3·s–1) range, the component,
after being immersed, has to be
completely stripped of attached air
bubbles so that the bubble formed by
leaking gas may be detected. The
10–5 Pa·m3·s–1 (10–4 std cm3·s–1) leakage
range is near the limit of detectability of
the bubble technique, although longer
waiting periods theoretically could obtain
higher sensitivity. Longer waiting periods
become impractical when the rate of
bubble evolution approaches the rate at
which tracer gas is dissolving in the test
fluid.
Specifying sensitivity much greater
than 10–5 Pa·m3·s–1 (10–4 std cm3·s–1)
makes bubble testing exceedingly difficult.
For instance, bubble testing could be used
at higher sensitivity by saturating the
immersion liquid with the tracer gas used
in leak testing. However, it would be
better to change to a different leak testing
method that is more effective at that
higher sensitivity. Bubble testing to detect
leaks greater than 10–2 Pa·m3·s–1
(10–1 std cm3·s–1) becomes difficult
because of rapid gas evolution and rapid
decay of pressure in the system under test.
However, difficulties in the less sensitive
test range are usually not so great as in
the more stringent sensitivity range.
Relation of Test Costs to
Sensitivity of Leak Testing
Leak testing instrumentation costs
increase as required test sensitivity
increases, as sketched in Fig. 2.5 The test
equipment investment for determining a
leakage rate of 10–4 Pa·m3·s–1
(10–3 std cm3·s–1) is negligible compared
with that for a sensitivity of
10–13 Pa·m3·s–1 (10–12 std cm3·s–1), whose
cost is 10 000 times higher. Even after a
test technique has been selected, raising
leak sensitivity requirements within this
technique will result in an increase in
measurement cost. This increase is usually
caused by greater complexity of leak tests
with increased sensitivity. Cost increases
become particularly drastic when the
required sensitivity is higher than the
optimum operating range shown in Fig. 1.
TABLE 2. Leak testing methods and techniques.
FIGURE 1. Ease of test operation as a function of leak testing
sensitivity.
Great
Ease of Operation
Optimum
operating range
Bubble solution
Ultrasonic/acoustic
Voltage discharge
Pressure
Ionization
Conductivity
Radiation absorption
Chemical based
Halogen detector
Radioisotope
Pressure change
Mass spectrometer
Low
High
Low
Leak Testing Sensitivity
12
Methods
Leak Testing
Techniques
immersion; film solution
sonic/mechanical flow; sound generator
voltage spark; color change
hydrostatic; hydropneumatic; pneumatic
photo ionization; flame ionization
thermal conductivity; catalytic combustible
infrared; ultraviolet; laser
chemical penetrants; chemical tracer gases
halide torch; electron capture; halogen
diode
krypton-85
absolute; reference; pressure rise; flow
measurement; pressure decay; volumetric
helium or argon; tracer probe location;
hooding total leakage; detector probe
location; sealed objects; residual gas
analyzer
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
Selection of Specific Leak
Testing Technique for
Various Applications6
Figure 3 provides a graphical guide to
selection of leak testing methods and
techniques for various applications. It
shows a decision tree with which the
choice of a leak testing method becomes a
step-by-step process. The selection
processes suggested by Fig. 3 serve as a
basic guide.5 Further consideration of
specific leak testing requirements may
suggest other methods or techniques for
test selection or cause the test engineer to
modify leak testing procedures. See also
Table 2. The final selection of the leak
testing method will typically be made
from perhaps only three or four possible
test methods. The special conditions
under which tests must be made can
become a major factor in this final test
selection.
The first question to be asked when
choosing the best leak testing method, or
technique of a method, is “Should this
test reveal the presence of a suspected
leak, or is its purpose to show the location
of a known leak?” The second question to
FIGURE 2. Effect of required sensitivity on leak detection
equipment cost.
50 000
Relative Leak Testing Equipment Cost (relative units)
Radioactive tracer techniques
5 000
Mass
spectrometer
500
Halogen heated anode
50
Bubble testing
5
10–4
10–7
10–10
10–13
(10–3)
(10–6)
(10–9)
(10–12)
Leakage Measurement Sensitivity,
Pa·m3·s–1
(std
cm3·s–1)
be answered is, “Is it necessary to measure
the rate of leakage at the specific leak?” If
leakage measurement is essential, use of
calibrated or reference leaks or other
means to provide quantitative leakage
measurement is required. In the decision
tree of Fig. 3, the first branch (or decision
point) answers the preceding questions
and determines if the purpose or
requirements of the test lead to the upper
branch of leak location only or to the
lower branch of leakage rate
measurements.
Basic Categories of Leak
Testing
Types of Fluid Media Used in Leak
Testing
Leak testing can be divided into three
main categories: (1) leak detection,
(2) leak location and (3) leakage
measurement. Each technique in all
categories involves a fluid leak tracer and
some means for establishing a pressure
differential or other means for causing
fluid flow through the leak or leaks.
Possible fluid media include gases, vapors
and liquids or combinations of these
physical states of fluid probing media.
Selection of the desired fluid probing
medium for leak testing depends on
operator or engineering judgment
involving factors such as: (1) type and size
of test object or system to be tested;
(2) typical operating conditions of test
object or system; (3) environmental
conditions during leak testing; (4) hazards
associated with the probing medium and
the pressure conditions involved in
testing; (5) leak testing instrumentation to
be used and its response to the probing
medium; (6) the leakage rates that must
be detected and the accuracy with which
measurements must be made; and (7)
compatibility of test probing medium
with test object and content (to avoid
corrosion etc.).
Gases and vapors are generally
preferred to liquid media where high
sensitivity to leakage must be attained;
however, liquid probing media are used
for leak testing in many specific
applications.
Selection of Tracer Gas
Technique for Leak
Location Only
As shown on the upper branch of the
decision tree of Fig. 3, tracer gas tests
whose purpose is leak location only can
be divided into a tracer probe technique
Introduction to Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
13
and a detector probe technique (see
Fig. 4).5 When choosing either technique,
it is important that leak location be
attempted only after the presence of a
leak has been ascertained. The tracer
probe technique is used when the test
system is evacuated and the tracer gas is
applied to the outside of the pressure
boundary of the test system. The detector
probe technique is selected when the test
system is pressurized with gases including
the tracer gas (if used) and the sniffing or
sampling of the leaking gas is being done
at atmospheric pressure in the ambient
air. This selection corresponds to the
second decision point in the upper
branch of the decision tree of Fig. 3.
FIGURE 3. Graphical decision tree for step-by-step selection of leak testing methods.
Halogen electron capture/halogen heated anode
Helium mass spectrometer
Infrared
Optical deflection
Gage response
Higher sensitivity
Chemical reaction
Inherent tracer
Gage in place
Detector
probe
Leak
location
Bubble
Helium mass spectrometer
Airborne ultrasonic
Argon mass spectrometer
Laser imaging
Residual gas analyzer
Acoustic emission
Infrared
Hydrostatic
Compare these factors in
choosing a leak testing
method or technique
Halogen heated anode
High voltage discharge
Pressurized
system
Gage response
Pressure measurement
Evacuated
system
Inherent detector
Lower equipment cost
Airborne ultrasonic
Tracer probe
Radioactivity
Helium mass spectrometer
Infrared
Back pressurizing
Inherent gage
Flow measurement
Evacuated
Multiple
sealed
Radioactivity
Mass spectrometer
Helium mass
spectrometer
Infrared
Leakage
rate
measurement
Sealed with
tracer
Dynamic testing
Low sensitivity
test run
after high
sensitivity test
Helium mass spectrometer
Halogen heated anode
Pressure change
Flow measurement
Halogen electron
capture/halogen
heated anode
Halogen
heated anode
Static testing
Halogen heated anode
Leak test
Radioactivity
Back pressuring
Pressure measurement
Air sealed
Bubble
Flow measurement
Low sensitivity
Inherent tracer
Open or single
sealed units
Gage in place
Dynamic testing
Optical deflection
High sensitivity
Halogen electron capture/
halogen heated anode
Infrared
Helium mass spectrometer
Static testing
Infrared
Bubble
Pressure measurement
Flow measurement
Leak to vacuum
Inherent tracer
Gage in place
Leak to atmosphere
14
Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
Factors Influencing Choice
between Detector Probe
and Tracer Probe Tests
One of the most difficult and important
decisions is the choice of which leak
testing method should be used. A correct
choice will optimize sensitivity, cost and
reliability of the leak testing procedure.
Choice of an incorrect test method makes
leak testing less sensitive and less reliable,
while adding to the difficulty of testing.
One simplified way to choose is to rank
various leak testing methods by means of
their leakage sensitivity. If this were
sufficient, the test engineer would only
need to decide what degree of sensitivity
is required and then to select the test
method from among those offering
adequate sensitivity for the specific test
application. However, each leak testing
technique can have a different test
sensitivity under different operating
conditions. For example, a mass
spectrometer leak detector is 10 000 times
more sensitive than a heated anode
halogen vapor detection instrument when
used for leak location in the tracer probe
leak location test of an evacuated vessel.
However, if these two instruments are
used for leak detection on a pressurized
test system, the halogen leak detector is
100 times more sensitive. The reason for
this apparent discrepancy becomes
obvious on close examination of the
FIGURE 4. Tracer gas probing for locating leaks with sensitive
electronic leak detection instruments; (a) tracer probe
technique; (b) detector probe technique.
(a)
Probe
System
under
test
Leak
detector
Source of
tracer gas
(b)
Probe
System
under
test
Source of
tracer gas
Leak
detector
operating characteristics of these two
instruments. The mass spectrometer is
designed for operation under vacuum
conditions, whereas the halogen leak
detector is designed for operation in air at
atmospheric pressure.
As another example, a helium mass
spectrometer leak detector may have a
leakage sensitivity of 10–12 Pa·m3·s–1
(10–11 std cm3·s–1) during routine leak
testing with dynamic leakage
measurement techniques. On very small
systems, this optimum sensitivity may be
increased to 10–15 Pa·m3·s–1
(10–14 std cm3·s–1), a gain of 1000×, by
using the static accumulation leakage
measurement technique. However, the
static leakage measurement technique is
not the standard method of using the
mass spectrometer leak detector.
Therefore, the last sensitivity stated above
is subject to some question. It must be
recognized that each method of leak
detection or measurement is usually
optimized for one particular type of leak
testing. Therefore, it can be a mistake to
compare sensitivities of various leak
testing methods under the same
conditions, if each test is not designed to
operate under these same conditions.
Leak Location Technique with
Detector Probe Operating at
Atmospheric Pressure
When testing a pressurized system that is
leaking into the atmosphere, the next
decision point is whether or not the
leaking fluid can be used as a tracer (this
decision point lies along the top branch
of the tree of Fig. 3). For example, most
refrigeration and air conditioning systems
are charged with a refrigerant gas
(refrigerant-22 or -134a) that is a
fluorocarbon to which the heated anode
halogen vapor detector is specifically
highly sensitive. When searching for leaks
in operating systems of this type, the
inherent tracer dictates the use of the
halogen leak testing method. Because of
potential environmental effects from
fluorocarbons, some current systems are
being charged with refrigerant-134a gas or
sulfur hexafluoride for use, respectively,
with modified residual gas analyzer
halogen leak detectors or electron capture
halogen leak detectors.
If the pressurized test system contains
ammonia gas, a chemical type of leak
detector might prove to be optimum. In
certain cases where the mass spectrometer
leak detector is to be used, the presence of
a specific gas (such as argon, helium or
neon) within the system provides an
excellent inherent tracer. Alternative
procedures involve pressurizing the test
system with such a tracer gas or a mixture
of air with tracer gas.
Introduction to Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
15
Some other methods for leak location
do not depend on the specific nature of
the leaking gas; among these are the
ultrasonic leak detector and bubble
testing. In some cases, the tracer gas
might be suitable for use with more than
one testing method, e.g., helium could be
used for bubble testing for large leaks or
for mass spectrometer testing for small
leaks or quantitative leakage
measurements.
The detector probe leak testing
methods, in order of increasing leak
sensitivity, time and costs, are ultrasonic,
bubble, chemical, pressure or flow gage
response, infrared gas detector, mass
spectrometer leak detector and halogen
vapor detector. These relative sensitivity
ratings apply for detector probes searching
with the detector inlet probe or sniffer
searching in air at atmospheric pressure.
These alternative leak test methods are
listed vertically at the right end of the top
branch of the decision tree of Fig. 3. The
lowest cost, highest speed, simplest leak
tests are at the bottom of this list. The
slower, more costly, higher sensitivity test
methods appear at the top of the list
shown to the right of the top branch of
the decision tree of Fig. 3.
Leak Location Technique with
Tracer Probe outside an Evacuated
System
When testing an evacuated system that
has in-leakage from the ambient
atmosphere or from a tracer probe, the
first consideration in selection of a test
method is whether there is an inherent
detector within the system. the inherent
detector might be a pressure gage of an
electronic type or, more desirably, a gage
that is specifically responsive to the
partial pressure of a specific tracer gas.
Vacuum systems often contain one or
more types of vacuum gages. In Fig. 3,
this point appears in the second main line
from the top, for tracer probe testing of
evacuated systems, and is labeled inherent
detector.
If a vacuum gage does not exist within
the evacuated system under test, other
test methods must be examined
individually to determine their limitations
and advantages for leak testing of this
system. The tracer probe leak testing
methods, in order of increasing leak
sensitivity, time and cost, are ultrasonic,
pressure change gage response, high
voltage electrical discharge, heated anode
halogen detector, infrared gas detector
and mass spectrometer helium leak
detector (highest in list). These methods
are listed vertically at the right end of the
second horizontal branch in Fig. 3.
The methods shown in the upper half
of Fig. 3 for leak location are those in
16
Leak Testing
primary or most common usage. Other
methods, such as those using radioactive
tracer gases, are not generally used
because of safety and other operating
problems associated with their use.
However, if none of the leak location
methods described for detector probe or
tracer probe leak tests in the preceding
discussion is satisfactory for a specific
application, more complicated leak testing
methods may be considered during
selection of an appropriate leak testing
method.
Selection of Technique for
Leakage Measurement
The lower half of the decision tree diagram
of Fig. 3 is a guide for step-by-step
selection of optimum techniques for
leakage measurements. Leakage
measurements can be divided into two
different types based on the nature of the
test objects whose leakage is to be
measured. The first decision is based on
the accessibility of test surfaces on the
pressure boundaries of the test object. Test
objects are classified by accessibility into
two groups.
1. Open units are accessible on both
sides of the pressure boundary, for
tracer probes or detector probes.
2. Sealed units are accessible only on
external surfaces.
The second category usually consists of
mass produced items such as transistors,
relays, ordnance components and sealed
instruments. In the lower portion of
Fig. 3, this choice is indicated first on the
decision path for leakage measurement.
Practical Measurement of
Leakage Rates with
Gaseous Tracers
Principles of Leakage
Measurement
All leak detection with tracer gases
involved their flow from the high pressure
side of a pressure boundary through a
presumed leak to the lower pressure side
of the pressure boundary. When tracer
gases are used in leak testing, instruments
sensitive to tracer gas presence or
concentration are used to detect outflow
from the low pressure side of the leak in
the pressure boundary. Where leak tests
involve measurements of change in
pressure or change in volume of gas
within a pressurized enclosure, the loss of
internal gas pressure or volume indicates
that leakage has occurred through the
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
pressure boundary (or temporary seals
placed on openings of the pressure
boundary). When evacuated or low
pressure test systems or components are
surrounded by higher pressure media such
as the earth’s atmosphere, or a hood or
test chamber containing gases at higher
pressures, leakage can be detected by loss
of pressure in the external chamber or by
rise in pressure within the lower pressure
system under test.
Classification of Techniques of
Leakage Measurement with Tracer
Gases
Leakage rate measurement techniques
involving the use of tracer gases fall into
two other classifications known as
(1) static leak testing and (2) dynamic leak
testing. In static leak testing, the chamber
into which tracer gas leaks and
accumulates is sealed and is not subjected
to pumping to remove the accumulated
gases. In dynamic leak testing, the
chamber into which tracer gas leaks is
pumped continuously or intermittently to
draw the leaking tracer gas through the
leak detector instrumentation, as sketched
in Fig. 5.5 The leakage rate measurement
procedure consists of first placing tracer
gas within or around the whole system
being tested. A pressure differential across
the system boundary is established either
FIGURE 5. Leakage measurement dynamic leak testing using
vacuum pumping: (a) pressurized system mode for leak
testing of smaller components; (b) pressurized envelope
mode for leak testing of larger volume systems.
(a)
Envelope
Leak detector
System
under
test
Source of tracer gas
(b)
Envelope
System
under
test
Leak detector
by pressurizing the one side of the
pressure boundary with tracer gas or by
evacuating the other side. The
concentration of tracer gas on the lower
pressure side of the pressure boundary is
measured to determine leakage rates.
Leakage Measurements of
Open Test Objects
Accessible on Both Sides
When test objects have pressure
boundaries accessible on both sides, the
second decision in the selection of a
leakage measurement test method is
whether the unit can or should be
evacuated during leak testing. This
decision will determine if the leak test is
performed with the tracer probe or
detector probe. If one side of the pressure
boundary can be evacuated so that
leakage occurs to vacuum and the leak
detector is placed in the vacuum system,
more sensitive leak testing will usually
result. In vacuum, the tracer gases can
reach the detector quickly, particularly
with dynamic tests in which the
evacuated test volume is pumped rapidly
and continuously. In this case, there is
little possibility of stratification of tracer
gases.
However, evacuation does not always
produce the most sensitive and reliable
leakage measurements. If the test volume
is extremely large, high pumping speed is
necessary to reduce response time. Such
auxiliary pumping will cause split flow,
thus reducing the amount of tracer gas
reaching the leak detector. This, in turn,
can reduce signal levels and leakage
sensitivity. Other restraints may prevent
evacuation of the test system to a
sufficiently low absolute pressure.
Conventional helium mass spectrometer
leak detectors, for example, should be
operated at vacuum levels of 0.1 Pa
(1 mtorr) or lower. Conventional helium
mass spectrometers can operate with
manifold vacuums of 2 Pa (20 mtorr) or
lower whereas counterflow helium mass
spectrometers can operate with manifold
vacuums of 10 Pa (0.1 torr) and higher.
The structure of the equipment under test
(particularly if thin walls not intended to
withstand external pressure are involved)
may prevent use of leakage rate
measurement techniques in which the
leak detector must operate within a
vacuum. In Fig. 3, the lowest branch
leading to the junction of the leak to
vacuum path and the leak to atmosphere
path represents the point of decision
discussed in this paragraph.
Source of tracer gas
Introduction to Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
17
Selecting Specific Method for Leak
Testing of Evacuated Test Units or
Systems
As indicated along the next-to-bottom
decision path at the center of Fig. 3, the
first approach to selecting leak test
methods for units that can be evacuated is
to determine whether or not there is an
inherent tracer in the test system while in
operation. For example, if in normal
operation the system under test contains
one of the specific tracer gases such as
helium or halogenated hydrocarbons, a
test method sensitive to that specific
tracer gas might be preferred. In this way,
considerable savings in test time and cost
can be realized if there is no need to fill
the system under test with a tracer gas.
If there is no inherent tracer gas within
the system under test, the next decision
step might be to determine if there is a
pressure or flow gage already present in
the evacuated system to be leak tested. If
so, this gage might be used for leakage
measurement in place of some additional
type of leak detector. This internally
available gage might be a simple vacuum
dial, thermocouple or ionization gage or,
in some fortunate cases, a mass
spectrometer that is incorporated into the
system as a part of its analytical
instrumentation or controls.
Consideration need not be limited to
those types of gages commonly used for
leak testing. Any gas concentration
measuring equipment that happens to be
available may be used for leakage
measurement and is accurate enough and
sensitive enough for the required results.
This decision point is that labeled gage in
place in the two bottom decision
pathways shown in Fig. 3.
Methods of Leakage
Measurement in Evacuated
Systems with No Inherent Tracer
If there is no inherent tracer or adequate
gage present within an evacuated test
system, other vacuum mode leak testing
methods must be considered. Methods for
leak testing of evacuated systems, in order
of increasing leak sensitivity and cost of
leak testing equipment, include gas flow
measurement, pressure change
measurement, heated anode halogen
vapor leak detection and mass
spectrometer helium leak detection. These
methods, listed vertically at the end of the
next-to-bottom decision line in Fig. 3,
should each be considered individually
and evaluated in terms of their
advantages and limitations. In most cases,
all of the possible leak testing methods
should be considered. Selection depends
on pertinent factors. For example, a more
sensitive leak testing method might
18
Leak Testing
involve higher initial costs for equipment
and test setups but, on the other hand, it
might result in great cost savings during
testing programs or provide greater
reliability in leak testing results.
Once the basic vacuum leak testing
method has been selected, a second
consideration involves selection between
static and dynamic test techniques. It is
usually preferable to perform leak tests
using a dynamic testing technique (tests
involving pumping of the vacuum system
throughout the test period). However,
static techniques of leakage rate
measurement should also be considered.
Static tests involving rise or loss in
pressure, or accumulation of tracer gases
over prolonged leak periods, are slower
than typical dynamic leak tests. However,
higher sensitivity can be achieved in static
tests if the volume under test is not
excessive; this may be worth the extra
effort.
Selection of Test Methods
for Systems Leaking to
Atmospheric Pressure
The choice of pressure mode testing
methods — i.e., for test systems leaking to
atmospheric pressure — should be made
by following the same type of decision
pattern as for leak testing of evacuated
systems. The decision path for this case
appears at the bottom of Fig. 3. The leak
testing methods applicable to testing of
systems leaking to atmosphere, in order of
increasing test sensitivity, are flow
measurement, pressure measurement (for
larger volume systems), immersion bubble
testing, infrared gaseous leak testing,
heated anode and electron capture
halogen leak testing, mass spectrometer
helium leak testing and leak testing using
radioactive tracer gases. A dynamic leak
testing method should be used wherever
possible. After various dynamic leak test
methods have been considered and those
whose limitations are unacceptable have
been rejected, a static leak testing method
should also be considered. Although a
static technique will increase leak testing
time, it will also increase leak testing
sensitivity.
Leak Testing to Locate
Individual Leaks
Leak testing for the purpose of locating
individual leaks is required when it is
necessary to detect, locate and evaluate
each leak; unacceptable leaks then can be
repaired and total leakage from a vessel or
system brought within acceptable limits.
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
Methods for detecting and locating
individual leaks are generally quantitative
only in the sense that the lower limit of
detectable leak size is determined by the
sensitivity of the leak detecting indicators
and test method used. Thus, only rather
crude overall leakage rate information
could be approximated by adding the
leakage rates measured for the leaks that
are detectable. Numerous different leak
detecting, locating and measuring
techniques and devices are available. The
selection of test equipment, tracer gas and
leak detection method is influenced by
the following factors: (1) size of the leaks
to be detected and located; (2) nature and
accuracy of leak test information required;
(3) size and accessibility of the system
being tested; (4) system operating
conditions that influence leakage;
(5) hazards associated with specific leak
location methods; (6) quantity of parts to
be tested; and (7) ambient conditions
under which leak location tests are
required to be carried out (wind or lack of
air circulation and stratification effects
can influence test sensitivity and
personnel).
Classification of Techniques for
Locating and Evaluating Individual
Leaks
Techniques for location and evaluation of
individual leaks can be categorized in
various ways, including by types of leak
tracer used in the detection, location and
possible measurements of individual leaks.
A primary classification is that between
the use of liquid tracers and the use of
more sensitive gaseous tracers. Leak
location techniques that depend on tracer
gas properties are listed below in general
categories, in order of increasing leak
testing sensitivity and complexity of test
methods:
1. leak location techniques independent
of any characteristic properties of the
tracer gas (use of candles, liquid and
chemical penetrants, bubble testing
and sonic or ultrasonic leak tests, for
example);
2. leak location techniques using tracer
gases with easily detectable physical or
chemical properties (gases with
thermal conductivities or chemical
properties differing from those of the
pressurizing gas, gaseous halogen
compounds and gases having
characteristic radiation absorption
bands in the ultraviolet or infrared
spectral ranges); and
3. leak location techniques involving the
use of tracer gases with atomic or
nuclear properties providing easily
detectable leak signals (helium and
other inert gases having specific
charge-to-mass properties that permit
their sensitive detection by mass
spectrometers and gaseous radioactive
isotopes detectable with particle
counters and radiation detectors).
Tables 3 and 4 list some typical leak
detection systems and give their leakage
sensitivities.
Techniques for Locating Leaks
with Electronic Detector
Instruments
Figure 4 shows arrangements of two basic
techniques for locating leaks with
electronic instruments that detect gas flow
or presence of specific tracer gases: (1) the
detector probe probe technique and
(2) the tracer technique. With either, it is
important that leak location pinpointing
be attempted only after the presence of a
leak has been ascertained. When choosing
between the pressure test technique and
the vacuum test technique, both of the
alternative techniques listed above must
be considered when the test object will
TABLE 3. Sensitivity limits of various methods of leak testing.
Method
Minimum Detectable
Leakage Rate
Pa·m3·s –1 (std cm3·s –1)
Mass loss
time limited
Ultrasonics
Penetrants
≤
Bubbles
Thermal conductivity
Halogen
0.05
10–4
10–5
10–6
10–10
(0.5)
(≤ 10–3)
(10–4)
(10–5)
(10–9)
Mass spectrometer
10–13
(10–12)
Comments
pressure change; generally limited to sizable leaks; good overall quantitative measure; no
information on leak location; time consuming
leak location only; fast; no cleanup; can detect from distance; large leaks only
simple to use; location only; may plug small leaks; requires cleanup
for leak location; fluids may plug small leaks; requires cleanup
simple; compact; portable; inexpensive; sensitive to various gases; operates in air
operates in air; sensitive (10–12 claimed with sulfur hexafluoride); portable; requires cleanup;
loses sensitivity with use; sensitive to ambient halide gases
most accurate for vacuum testing; expensive; relatively complex; not as portable as halogen
detectors; much less sensitive when used in detector probing
Introduction to Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
19
withstand either pressure or vacuum. If a
satisfactory choice of one technique has
been made, it is a good idea to compare it
with a satisfactory choice of the other
technique, to see if reduced cost or an
easier test method might be possible.
The detector probe leak location
technique is used when the system under
test is pressurized and testing is done at
ambient atmospheric pressure. The tracer
probe technique is usually used when the
system under test is evacuated and the
tracer gas comes from outside this system.
The tracer probe technique is usually the
most rapid test because the tracer gas
travels more rapidly in vacuum and so
reaches the leak detector in a shorter time.
On the other hand, a higher pressure
differential can be used with the detector
probe.
Coordinating Overall Leakage
Measurements with Leak Location
Tests
Leakage rate measurement techniques do
not provide information on the number
and locations of individual leaks. The
TABLE 4. Relative ultimate leakage sensitivities of leak
testing methods under ideal conditions with very high
concentrations of tracer gases.a
Test Method
Minimum Detectable
Leakage Rate
Pa·m3·s–1
(std cm3·s–1)
—— b
——
—— c
10 –2
10–3
10–3
10–3 to 10–4
10–3 to 10–4
10–3 to 10–4
10–3 to 10–4
10–4
10–4 to 10–5
10–4 to 10–5
10–5
6 × 10–5 to
6 × 10–7
Hydrogen Pirani
10–7
Hot filament ionization gage
10–7 to 10–8
Mass spectrometer detector probe 10–6 to 10–7
Halogen diode detector
10–7 to 10–9
Hydrogen bubbles in alcohol
5 × 10–7
Paladium barrier detector
10–8 to 10–9
Mass spectrometer envelope test 10–10
Radioactive isotopes
10–9 to 10–13
Liquid pressure drop
Gas pressure drop
Pressure rise
Ultrasonic leak detector
Volumetric displacement d
Gas discharge
Ammonia and phenolphthalein
Ammonia and bromocresol purple
Ammonia and hydrochloric acid
Ammonia and sulfur dioxide
Halide torch
Air bubbles in water
Air and soap or detergent
Thermal conductivity
Infrared
a.
b.
c.
d.
—— b
——
—— c
(10 –1)
(10–2)
(10–2)
(10–2 to 10–3)
(10–2 to 10–3)
(10–2 to 10–3)
(10–2 to 10–3)
(10–3)
(10–3 to 10–4)
(10–3 to 10–4)
(10–4)
(6 × 10–4 to
6 × 10–6)
(10–6)
(10–6 to 10–7)
(10–5 to 10–6)
(10–6 to 10–8)
(5 × 10–6)
(10–7 to 10–8)
(10–9)
(10–8 to 10–12)
Numbers not to be used as guides in practical leak testing.
Depends on volume tested and pressure range of gage.
Depends on volume tested.
Gas type flow meters.
20
Leak Testing
latter can only be determined by leak
location test techniques. However, use of
the leak location techniques alone cannot
give reliable assurance that no leaks exist
or that tests have revealed all leaks that
exist. Without prior assurance that leaks
do exist, leak location test techniques
become arbitrary in application.
In practice, preliminary leakage testing
is often done first by less sensitive
methods to permit detection, location and
rectification of gross leaks. Next, the
operator can determine if any additional
leakage exists by an overall leakage
measurement of the entire test vessel,
system or component. Then each
individual leak should be discovered by
sensitive leak location techniques and
repaired if feasible, until all detectable
leak locations have been identified and
their leaks rectified. For final assurance
that the test object or system meets
leakage specification requirements, it may
be necessary to repeat the overall leakage
rate measurement to determine whether
the total leakage rate falls within the
acceptable limits.
Training of Leak Testing
Personnel7
Because of the many leak testing
techniques and the multiple variations of
each, leak testing could require more
training and knowledge than any of the
other nondestructive testing methods.
Successful execution of many of these
techniques by inspection personnel is
highly dependent on knowledge and skill.
Nevertheless, there are fewer instruction
and training materials available for leak
testing than for other methods.
Leak testing may be divided into four
methods: bubble testing, pressure change
testing, halogen diode leak testing and
mass spectrometer leak testing (see
Table 2), to which may be added acoustic
methods. The outline for the Level I leak
testing methods course in Recommended
Practice No. SNT-TC-1A expands this list of
four methods to a total of 12 techniques.8
The 34 variations in Table 2 reveal the
complex nature of leak testing and may
also be the reason why such a small
percentage of ASNT membership is
qualified to Level III in the leak test
method. At Level I, proficiency in one or
two techniques is possible, but it would
be very difficult to meet the training and
experience guidelines that are
recommended by ASNT for more than
two or three techniques. A brief listing for
each technique may make you aware of
your weaknesses. Variations of each
technique may require familiarity with
different test equipment and tracers.
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
Many inspection people are also
confused, when choosing a technique, by
the disadvantages and limitations in
sensitivity for each technique.
Inspection personnel often have
difficulty understanding how extremely
small some leaks are that they will try to
find. This also makes it difficult to realize
that some leaks may be temporarily sealed
by foreign material such as oil, grease,
water even cleaning solvents or even
moisture in air. Improper handling after
cleaning may temporarily prevent
location of leaks that will reappear at a
later time. A comparison of leakage rates
in three different ways (Table 5) may help
to visualize the size.
When leak testing is performed with
equipment capable of locating and
measuring leaks smaller than
10–9 Pa·m3·s–1 (10–8 std cm3·s–1), tracer gas
permeation through the test object
materials of construction may appear as a
leak indication several seconds to hours
after application of the tracer. This may
require a knowledge of those materials
that allow permeation by the tracer being
used.
Many Level II or III inspection
personnel establish reject specifications
that are unrealistically small with respect
to the expected life of the product being
tested. As a result, many tested objects
with leaks that are 10 to 100 times smaller
than an acceptable level are rejected for
repair or destruction. This creates
unnecessary cost and loss of profits. Some
examples of leaks that may affect certain
products are as follows: chemical process
equipment, 10–2 to 10–1 Pa·m3·s–1 (10–1 to
1 std cm3·s–1); torque converter, 10–4 to
10–5 Pa·m3·s–1 (10–3 to 10–4 std cm3·s–1);
beverage can end, 10–6 to 10–7 Pa·m3·s–1
(10–5 to 10–6 std cm3·s–1); vacuum process
system, 10–7 to 10–8 Pa·m3·s–1 (10–6 to 10–7
std cm3·s–1); integrated circuit package,
10–8 to 10–9 Pa·m3·s–1 (10–7 to
10–8 std cm3·s–1); pacemaker,
10–10 Pa·m3·s–1 (10–9 std cm3·s–1).
Another reason training must be
emphasized is that many leak testing
hazards may exist that cause injury to
inspection personnel, damage to test
equipment or damage to the product
being tested. The following examples
illustrate numerous hazards:
flammable/toxic solvents for cleaning,
flammable/toxic/explosive tracers,
asphyxiation by vapors or tracer gases,
access difficult on large objects,
pneumatic and hydrostatic pressure,
radioactive tracer gases, compressed gas
cylinders/regulators and structural stress.
To summarize the need for leak testing
methods training, there are eleven reasons
to expand this training: choice of many
techniques, sensitivity of various
techniques, advantages and limitations of
each technique, dependence of
techniques on testing skills and
experience, leakage location versus
measurement, factors affecting
measurement accuracy, employers’ cutting
cost by hiring entry level people and
minimizing training time, hazards to
personnel and products, few courses
available that offer skills training, limited
available training materials and the small
number of qualified Level III personnel.
TABLE 5. Comparison of leak rates.
Measurementa
std cm3·s–1
Equivalentb
10–2
10–3
10–4
10–5
10–6
10–7
10–8
10–9
10–10
10–11
10–12
1
1
3
1
1
1
3
1
1
1
1
std
std
std
std
std
std
std
std
std
std
std
cm3/10 s
cm3/100 s
cm3/h
cm3/3 h
cm3/24 h
cm3/2 wk
cm3/yr
cm3/3 yr
cm3/30 yr
cm3/300 yr
cm3/3000 yr
Bubble
Equivalentb,c
steady stream
10 s–1
1 s–1
0.1 s–1
—— d
—— d
—— d
—— d
—— d
—— d
—— e
a. 1 std cm3·s–1 = 0.1 Pa·m3·s–1.
b. Approximate.
c. Assuming bubble of 1 mm3 (6.1 × 10–5 in.3)
volume.
d. Bubbles too infrequent to observe or partially
dissolved.
e. Smallest detectable leak by mass spectroscopy.
Introduction to Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
21
PART 3. History of Leak Testing9
According to modern accounts, making a
vacuum was generally considered
impossible until the mid-1600s. However,
leaks have concerned technologists for
thousands of years.
Despite the importance of leaks for
ship construction, nothing on methods of
caulking is to be found in reference works
in the history of ancient technology. Leak
testing, up to the era of vacuum,
depended solely on the eye and was so
commonplace as to escape attention. At
any event, references to leak testing are
hard to find until well into the 1800s.
Ruhmkorff and Tesla Coils
as Leak Detector
Although Nollet in Paris observed the
electric discharge in an exhausted vessel
in 1740, it was not until a century later
that substantial investigation of this low
pressure discharge took place. Michael
Faraday, in 1831, had enunciated the
principle of the induction coil and had
studied discharges in gases by 1839.
By about 1850, Ruhmkorff and others
had made substantial improvements in
Faraday’s coil. Presumably, development
of the Ruhmkorff induction coil and the
Tesla coil greatly facilitated investigation
of the high voltage vacuum discharge.
By 1859, there were reports by Gassiot
and others of the changing nature of the
discharge with pressure. Moreover, it was
observed that the color of the discharge
depended on the gas in the discharge tube
as well as on the pressure.
It seems likely that, soon after 1860,
high voltage was applied to glass systems
to determine the presence of leakage.
Besides being sensitive to pressure and
chemistry, the discharge tends to enter
the system through the leaks, the air in
the leak offering a low resistance path.
Nineteenth Century Leak
Testing
In previous centuries, in the absence of
precise instrumentation for measurement
of flow, pressure or chemical
concentrations, leak testing had to rely on
methods that emphasized detection of
gross leak by making the leaking
22
Leak Testing
substance more conspicuous and hence
making the leakage easier to find.
Natural Gas Pipe Leak Testing
In the 1880s, inventor George
Westinghouse patented a means of
detecting leakage of fossil gas through gas
pipelines. The idea was essentially to
encase or sheathe one pipe within
another. The zone between the two pipes
could then be monitored to detect gas
leaking from the interior pipe. As
principal owner of utilities and gas
delivery systems based in western
Pennsylvania, Westinghouse had a strong
commercial interest in leak testing.10
Smoke Tracer
A leak detection device has a role in the
story “A Scandal in Bohemia” in the
Adventures of Sherlock Holmes (1892) by
Arthur Conan Doyle. Sherlock Holmes
assumes a disguise and gains admittance
to a woman’s lodgings to recover love
letters compromising to his client. At a
prearranged moment, Dr. Watson throws
a smoke bomb, called a plumber’s smoke
rocket, in through a window and calls
“fire.” The lady promptly goes to rescue
the love letters, thereby revealing their
hiding place. Not rockets at all in the
modern sense, smoke bombs were used by
plumbers who would ignite and put them
in piping and ductwork so that smoke
would reveal leaks.
Pressure Gages
After the invention of the high voltage
sparker in the mid-1800s, no advances in
leak detection methods are documented
until after the turn of the century. In
1906, Pirani described his hot wire
manometer, the well known Pirani gage.
The resistance of an electrically heated
wire was measured continuously to
determine the temperature of the wire,
the temperature increasing with decrease
in pressure.
That same year, W. Volge published a
description of a hot wire manometer
known as the thermocouple gage in which
the temperature of the wire was indicated
by the output of a thermocouple welded
to the wire. Both the Pirani and
thermocouple gages are affected by the
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
residual gases in the vacuum. Accordingly,
exposing the system to a gas such as
hydrogen, or painting suspected leakage
points with liquids such as alcohol or
acetone, results in changes of gage output
when a leak has actually been covered.
Hot Cathode Gage
There are many gages that can be used as
leak detectors because their outputs are
functions of the system residual gases. But
the most sensitive is the hot cathode
ionization gage because it measures the
lowest pressures. This was described (but
not illustrated) by Oliver E. Buckley in
1916.
It is to be noted that Adolf von Bäyer,
in 1909, used both a diode and a triode to
measure ionization currents but did not
suggest their use as pressure gages.
McLeod invented the gage (named after
him) in 1874. This gage, and several other
gages earlier than the Buckley ionization
gage, are not used for leak testing either
because they do not have a continuous
output or because they are difficult to
manufacture and/or use.
Helium Mass Spectrometer
Leak Detector
Developed in 1910, the mass spectrometer
had as its first achievement the positive
confirmation of the existence of isotopes,
specifically those of neon. The instrument
was improved rapidly so that it became a
tool for precision determination of
particle mass and relative isotopic
abundance. Perhaps its most familiar
application is the quantitative and
qualitative analysis of chemical
compounds and mixtures. However, one
of the earliest and presently the largest
single application of mass spectrometers is
that of the location and measurement of
extremely fine leaks.
During the Second World War, the
Manhattan Project had been formed in
the United States Corps of Engineers to
build atomic bombs. An essential part of
its assignment was to separate substantial
quantities of radioactive uranium-235
from uranium-238, with which it occurs
in ores. One approach to this separation
was embodied in the diffusion plant built
in Oak Ridge, Tennessee.
The plant was to operate on uranium
in the form of uranium hexafluoride (UF6)
in the vapor state, and it was realized
early on that the process equipment
would have to be free from leaks. The
lowest pressure in the system was to be
about 10 Pa (0.1 atm), so that loss of
vacuum was not a concern. First of all,
there was the possible outflow of uranium
hexafluoride, which is corrosive and
poisonous and which would include loss
of precious uranium-235. But the real fear,
amounting to a nightmare, was the
possible inflow of moist air.
The Oak Ridge plant was to consist of
acres of diffusion barrier, and the barrier
was to be a membrane containing billions
of holes of diameters less than 10 nm
(4 × 10–7 in.), the mean free path of
uranium hexafluoride being about
100 nm (4 × 10–6 in.). Moist air would
react with uranium hexafluoride to form
uranium oxide in the form of finely
divided powder. Conceivably, in the first
day of operation of the plant, this powder
could clog all the barrier pores, and the
most expensive and important war project
the United States had ever undertaken
would be unsuccessful.
Consequently, a subgroup was set up to
determine or develop a suitable hole
detection device. The group was headed
by Robert B. Jacobs, who was given the
task of developing the most sensitive
detection system he and his group could
devise.
A number of approaches were tried,
including the use of a variety of trapped
vacuum gages and an optical
spectrometer, all of which lacked either
the necessary sensitivity and/or selectivity.
Jacobs was aware that A.O.C. Nier of the
University of Minnesota, Minneapolis,
was doing work with a relatively simple
type of mass spectrometer of his own
design — a 60 degree sector instrument.
Nier had used his spectrometer to obtain
the first samples of uranium-235
separated from uranium-238.
At Jacobs’ behest, Nier devised a leak
detector, based on a simplified mass
spectrometer gas analyzer, that used a hot
filament cathode and was designed to
detect helium as a search gas. Helium had
been chosen as the leak probe gas because
of its very low concentration — one part
per 200 000 — in atmospheric air. In
theory the spectrometer was selective but
actually at the time there were some
interferences.
Leak Testing for Efficiency
Improvement
The helium leak detector is by far the
most sensitive device of its kind. In 1945,
its sensitivity was in the neighborhood of
10–7 Pa·m3·s–1 (10–6 std cm3·s–1). This was
100 or more times more sensitive than an
ionization gage, the next most sensitive
device. Today’s mass spectrometer leak
detectors can detect flows of
10–12 Pa·m3·s–1 (10–11 std cm3·s–1), i.e.,
leaks 105 times smaller than the original
models.
While waiting for the mass
spectrometer’s delivery, a number of
Introduction to Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
23
accessories essential to the reliable use of
the instrument were being developed.
These included calibrated leaks of the
flattened tube type, portable setups for
preparing helium-air mixtures of low,
known helium concentration, vacuum
tight metallic quick connects and pump
stations.
When the first few mass spectrometers
finally arrived, it was found that each
spectrometer was made of glass and
included a glass mercury high vacuum
pump. The electron emitting filament was
fused into the glass tube. Whenever a
filament burned out, an expert
glassblower was required to crack the
filament out of the tube and fit in another
with precisely the right orientation.
Nevertheless, the units were tested for
sensitivity (about 1 part helium in 100
000 parts of air mixture) and sent to
project contractors such as Chrysler
Corporation.
Although mass spectrometers were
typically made of glass then, the leak
testing personnel at manufacturing plants
during the war were continually burning
out the filaments and accidentally
breaking the glass tubes. After being
chided several times, they finally
threatened to quit.
Jacobs was asked to resolve this crisis
and came up with the idea of an all metal
system that included the spectrometer
tube. Nier’s reaction was negative because
in his experience metal had never been
used for the mass spectrometer tube and
he could think of a number of reasons
why it wouldn’t work. At Jacobs’ urging,
however, the project was undertaken by
Nier and his University of Minnesota
group.
In a few months, a first model was
constructed and worked as well as the
original glass one. Moreover, the filament
was now mounted into a standard glass
male taper. It was a relatively simple job
to align this in a companion metal taper
mounted on the metal mass spectrometer
tube and seal it with vacuum wax. And so
the Nier-Keller-General Electric leak
detector (Fig. 6) was born. The Nier-Keller
prototype was given to General Electric to
reengineer and manufacture, and General
Electric supplied all the detectors used for
diffusion plant testing.
The diffusion plant equipment was
designed and constructed along lines laid
down by Jacobs’ group, to facilitate leak
testing. The plant worked, substantial
quantities of uranium-235 were prepared,
and the leak detector successfully
performed its mission. However, rumor
had it that leak tightness of the plant did
not have to be as extreme as originally
thought.
Immediately after the war, leak
detectors were being offered to the public
24
Leak Testing
and found immediate and widespread
application to the testing of electron tubes
and to atomic work, the age of the
particle accelerator having begun.
Contemporary Leak
Detectors
In the years since 1945, the helium
detector has undergone somewhat
spectacular improvement, although the
FIGURE 6. Nier’s helium mass spectrometer leak detector
(circa 1944): (a) schematic; (b) photograph.
(a)
Emission regulator connection
Gas inlet
Focus plates
Baffles
Iron pole piece
Block
magnet
Baffles
Electron tube
mp
u
p
To
Collector slit
Suppressor plate
Collector plate
Collector rod
Input resistor
Amplifier
connection
(b)
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
change may be typical of what happens
with any new instrument. In 1945, the
sensitivity for helium was about 10 parts
helium in 1 million parts of air mixture,
or about 103 Pa·m3·s–1 (104 std cm3·s–1). By
the late 1950s, this figure had gone to
about 10–9 or 10–10 Pa·m3·s–1 (10–8 or
10–9 std cm3·s–1). For a number of years
now, commercial units have been
providing sensitivities better than
10–11 Pa·m3·s–1 (10–10 std cm3·s–1). The
equivalent parts-per-million figure is
100 to 10 nL·L–1. Obviously, helium in air
can now easily be detected.
Size has been reduced even though an
extra mechanical pump for roughing has
been included in the instrument cabinet.
In recent years, several mobile units have
been made available. The weight
reduction in these units is accomplished
in part by eliminating the cold trap and
by using a small mechanical pump that
acts as both a diffusion pump backer and
a test line roughing pump.
The Oak Ridge detector had manually
controlled valves. Operator error
frequently resulted in admission of
atmospheric pressure to the unit, with
attendant casualties to the mass
spectrometer filament, the pump oil and
the system. Models in the 1990s
automatically monitor gas admission to
the detector and give automatic, digital
readout of the leak rate of the defect
being probed. Some units require only the
depressing of a button to start the
detecting task. So-called industrial leak
testing systems are available for testing
mass produced components. The operator
needs only to place the test object into a
rack and press a start button. The system
operates automatically and flashes a go or
no-go signal at the end of the test.
Helium mass spectrometer leak
detectors became commercially available
in the United States in the late 1940s. The
versatility of mass spectrometer
instruments has led to a wide variety of
applications. Presently, thousands of these
sensitive leak detectors are in use
throughout the world. Leak detectors are
found in almost every university,
industrial or government physics
laboratory.
Thanks to these historic developments,
a tremendous amount of time has been
saved in leak testing operations. Whereas
days and even weeks were spent in
finding leaks in laboratory high vacuum
setups, the helium detector made it
possible to locate the leaks in hours or
minutes. Nier will probably be most
remembered in the annals of physics for
his work in mass spectroscopy but the
scientific world is more in his debt for the
leak detector.
Introduction to Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
25
PART 4. Units of Measure for Nondestructive
Testing
Origin and Use of the SI
System
In 1960 the General Conference on
Weights and Measures devised the
International System of Units. Le Systeme
Internationale d’Unites (SI) was designed so
that a single set of interrelated
measurement units could be used by all
branches of science, engineering and the
general public. Without SI, this
Nondestructive Testing Handbook volume
could have contained a confusing mix of
Imperial units, obsolete centimeter-gramsecond (cgs) metric system version units
and the units preferred by certain
localities or scientific specialties.
SI is the modern version of the metric
system and ends the division between
metric units used by scientists and metric
units used by engineers and the public.
Scientists have given up their units based
on centimeter and gram and engineers
made a fundamental change in
abandoning the kilogram-force in favor of
the newton. Electrical engineers have
retained their amperes, volts and ohms
but changed all units related to
magnetism. The main effect of SI has been
the reduction of conversion factors
between units to one (1) — in other
words, to eliminate them entirely.
Table 6 lists seven base units. Table 7
lists derived units with special names.
Table 8 gives examples of conversions to
SI units. In SI, the unit of time is the
second (s) but hour (h) is recognized for
use with SI.
For more information, the reader is
referred to the information available
through national standards organizations
TABLE 6. Base SI units.
Quantity
Length
Mass
Time
Electric current
Temperaturea
Amount of substance
Luminous intensity
Unit
Symbol
meter
kilogram
second
ampere
kelvin
mole
candela
a. Kelvin can be expressed in degrees celsius
(°C = K – 273.15).
26
Leak Testing
m
kg
s
A
K
mol
cd
and specialized information compiled by
technical organizations.11-13
Multipliers
Very large or very small numbers with
units are expressed by using the SI
multipliers, prefixes of 103 intervals
(Table 9) in science and engineering. The
multiplier becomes a property of the SI
unit. For example, a millimeter (mm) is
0.001 meter (m). The volume unit cubic
centimeter (cm3) is (0.01)3 or 10–6 m3.
Unit submultiples such as the centimeter,
decimeter, dekameter (or decameter) and
hectometer are avoided in scientific and
technical uses of SI because of their
variance from the 103 interval. However,
dm3 and cm3 are in use specifically
because they represent a 103 variance.
TABLE 7. Derived SI units with special names.
Quantity
Frequency (periodic)
Force
Pressure (stress)
Energy
Power
Electric charge
Electric potentialb
Capacitance
Electric resistance
Conductance
Magnetic flux
Magnetic flux density
Inductance
Luminous flux
Illuminance
Plane angle
Radioactivity
Radiation absorbed dose
Radiation dose equivalent
Solid angle
Time
Volumec
Units
hertz
newton
pascal
joule
watt
coulomb
volt
farad
ohm
siemens
weber
tesla
henry
lumen
lux
radian
becquerel
gray
sievert
steradian
hour
liter
Symbol
Hz
N
Pa
J
W
C
V
F
Ω
S
Wb
T
H
lm
lx
rad
Bq
Gy
Sv
sr
h
L
Relation
to Other
SI Unitsa
1·s–1
kg·m·s–2
N·m–2
N·m
J·s–1
A·s
W·A–1
C·V–1
V·A–1
A·V–1
V·s
Wb·m–2
Wb·A–1
cd·sr
lm·m–2
1
1·s–1
J·kg–1
J·kg–1
1
60 s
dm3
a. Number one expresses dimensionless relationship.
b. Electromotive force.
c. The only prefixes that may be used with liter are milli (m) and micro (µ).
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
LT.01 LAYOUT 11/8/04 2:12 PM Page 27
Note that 1 cm3 is not equal to 1/100 m3.
Also, in equations, submultiples such as
centimeters (cm) or decimeters (dm)
should be avoided because they disturb
the convenient 103 or 10–3 intervals that
make equations easy to manipulate.
In SI, the distinction between upper
and lower case letters is meaningful and
should be observed. For example, the
meanings of the prefix m (milli) and the
prefix M (mega) differ by nine orders of
magnitude.
approved for use. The liter is a special
name for cubic decimeter (1 L = 1 dm3 =
10–3 m3). Only the milli (m) and micro (µ)
prefixes may be used with liter.
The fundamental units of time,
temperature, pressure and volume are
expressed every time a leakage is
measured.
Units for Measurement of
Radioactive Tracer Gases
The pascal (Pa), equal to one newton per
square meter (1 N·m–2), is used to measure
pressure, stress etc. It is used in place of
units of pound force per square inch
(lbf·in.–2), atmosphere, millimeter of
mercury (mm Hg), torr, bar, inch of
mercury (in. Hg), inch of water (H2O) and
other units (see Table 10). The text must
indicate whether gage, absolute or
differential pressure is meant. Negative
pressures might be used in heating duct
technology and in vacuum boxes used for
bubble testing, but in vacuums as used in
tracer leak testing absolute pressures are
used.
The original curie was simply the
radiation of one gram of radium.
Eventually all equivalent radiation from
any source was measured with this same
unit. The original roentgen was the
quantity of radiation that would ionize
one cubic centimeter of air to one
electrostatic unit of electricity of either
sign. It is now known that a curie is
equivalent to 3.7 × 1010 disintegrations
per second and a roentgen is equivalent
to 258 microcoulomb per kilogram
(258 µC.kg–1) of air. This corresponds to
1.61 × 1015 ion pairs per kilogram of air
that has absorbed 8.8 millijoule (mJ) or
0.88 rad.
In SI, radiation units have been given
established physical foundations and new
names where necessary. The unit for
radioactivity (formerly curie) is the
becquerel (Bq), defined as one
disintegration per second.
Volume
Derived SI Units
SI Units for Leak Testing
Pressure
(m3)
The cubic meter
is the only volume
measurement unit in SI. It takes the place
of cubic foot, cubic inch, gallon, pint,
barrel and more. In SI, the liter (L) is also
Gas Quantity. Pascal cubic meter (Pa·m3).
The quantity of gas stored in a container
or which has passed through a leak is
described by the derived SI unit of pascal
TABLE 8. Examples of conversions to SI units
Quantity
Measurement in Non-SI Unit
Multiply by
To Get Measurement in SI Unit
square inch (in.2)
645
square millimeter (mm2)
angstrom (Å)
0.1
nanometer (nm)
inch (in.)
25.4
millimeter (mm)
Energy
British thermal unit (BTU)
1.055
kilojoule (kJ)
calorie (cal), thermochemical
4.184
joule (J)
0.293
watt (W)
British thermal unit per hour (BTU·h–1)
Specific heat
British thermal unit per pound
4.19
kilojoule per kilogram per kelvin (kJ·kg–1·K–1)
–1
–1
per degree Fahrenheit (BTU·lbm ·°F )
Force (torque, couple)
foot-pound (ft-lbf)
1.36
joule (J)
Force or pressure
pound force per square inch (lbf·in.–2)
6.89
kilopascal (kPa)
Frequency (cycle)
cycle per minute
1/60
hertz (Hz)
Illuminance
footcandle (ftc or fc)
10.76
lux (lx)
Luminance
candela per square foot (cd·ft–2)
10.76
candela per square meter (cd·m–2)
candela per square inch (in.·ft–2)
1 550
candela per square meter (cd·m–2)
footlambert
3.426
candela per square meter (cd·m–2)
lambert
3 183 (= 10 000/π) candela per square meter (cd·m–2)
Radioactivity
curie (Ci)
37
gigabecquerel (GBq)
Ionizing radiation exposure roentgen (R)
0.258
millicoulomb per kilogram (mC·kg–1)
Mass
pound (lbm)
0.454
kilogram (kg)
Temperature (difference)
degree fahrenheit (°F)
0.556
degree celsius (°C)
Temperature (scale)
degree fahrenheit (°F)
(°F – 32)/1.8
degree celsius (°C)
(°F – 32)/1.8) + 273.15 kelvin (K)
Area
Distance
Introduction to Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
27
cubic meter, the product of pressure and
volume. To be strict, the temperature
should be specified for the gas quantity or
leakage measurement to define the gas
quantity (sometimes loosely described as
the mass of gas) more precisely. Often, gas
quantity is defined for standard
temperature and pressure, typically the
standard atmospheric pressure of 100 kPa
(1 atm) and a temperature of 0 °C (32 °F).
Temperature corrections are usually
required if temperature varies significantly
during leak testing. However, small
changes in temperature may sometimes
be insignificant compared with many
orders of magnitude of change in gas
pressure or leakage quantity.
Gas Leakage Rate. Pascal cubic meter per
second (Pa·m3·s–1). The leakage rate is
defined as the quantity (mass) of gas
leaking in one second. The unit in prior
use was the standard cubic centimeter per
second (std cm3·s–1). Use of the word
standard in units such as std cm3·s–1
requires that gas leakage rate be converted
to standard temperature and pressure
conditions (293 K and 101.325 kPa), often
even during the process of collecting data
during leakage rate tests. Leakage rates
given in SI units of Pa·m3·s–1 can be
converted to units of std cm3·s–1 at any
time by simply multiplying the SI leakage
rate by 10 or (more precisely) by 9.87.
Gas Permeation Rate. Pascal cubic meter
per second per square meter per meter
(Pa·m3·s–1)/(m2·m–1). Permeation is the
leakage of gas through a (typically solid)
yotta
zetta
exa
peta
tera
giga
mega
kilo
hectoa
deka (or deca)a
decia
centia
milli
micro
nano
pico
femto
atto
zepto
yocto
Symbol
Multiplier
Y
Z
E
P
T
G
M
k
h
da
d
c
m
µ
n
p
f
a
z
y
1024
1021
1018
1015
1012
109
106
103
102
10
10–1
10–2
10–3
10–6
10–9
10–12
10–15
10–18
10–21
10–24
a. Avoid these prefixes (except in dm3 and cm3) for
science and engineering.
28
Leak Testing
(1)
1.0
std cm 3⋅ s –1
cm 2 ⋅ cm –1
≅
0.1
Pa ⋅ m 3⋅ s –1
m 2 ⋅ m –1
Rounding
Many tables and graphs were obtained
from researchers and scientists who did
their work in the English system. In the
TABLE 10. Conversion factors for pressure
values.
To Convert
From
To
pascal (Pa)
lbf·in.–2
kg·mm–2
atm
in. Hg
torr
pound per square
inch (lbf·in.–2)
TABLE 9. SI multipliers.
Prefix
substance that is not impervious to gas
flow. The permeation rate is larger with an
increased exposed area, a higher pressure
differential across the substance
(membrane, gasket etc.) and is smaller
with an increasing thickness of permeable
substance. In vacuum testing, the pressure
differential is usually considered to be one
atmosphere (about 100 kPa). One
sometimes finds units of permeation rate
where the gas quantity is expressed in
units of mass and where the differential
pressure is expressed in various units.
Equation 1 expresses an equivalence for
conversion of measurements:
Pa
kg·mm–2
atm
in. Hg
torr
kilogram per square Pa
millimeter
lbf·in.–2
–2
(kg·mm )
atm
in. Hg
torr
Multiply by
1.4504
1.0197
9.8692
2.9530
7.5006
×
×
×
×
×
10–4
10–7
10–6
10–4
10–3
6.8948 × 103
7.0307 × 10–4
6.8046 × 10–2
2.0360
51.715
9.8066
1.4223
96.784
2.8959
7.3556
× 105
× 103
× 103
× 104
atmosphere (atm)
Pa
1.01325 × 105
lbf·in.–2
14.696
kg·mm–2
1.0332 × 10–2
in. Hg
29.921
torr
760.0
inch mercury
(in. Hg)
Pa
lbf·in.–2
kg·mm–2
atm
torr
3.3864
4.9115
3.4532
3.3421
25.40
×
×
×
×
103
10–1
10–4
10–2
torr
Pa
lbf·in.–2
kg·mm–2
atm
in. Hg
1.3332
1.9337
1.3595
1.3158
3.9370
×
×
×
×
×
102
10–2
10–5
10–3
10–2
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
conversion, some numbers have been
rounded drastically but some were left as
irrational numbers, especially where
quotes were made to specific entries.
Quantitative Description of
Leakage Rates
The significant quantitative measurement
resulting from leak testing is the
volumetric leakage rate or mass flow rate
of fluid through one or more leaks.
Leakage rate thus has dimensions
equivalent to pressure times volume
divided by time. The units used previously
for volumetric leakage rate were standard
cubic centimeter per second (std cm3·s–1).
The Nondestructive Testing Handbook uses
the international standard SI
nomenclature. In SI units, the quantity of
gas is measured in units of pascal cubic
meter (Pa·m3). The leakage rate is
measured in pascal cubic meter per
second (Pa·m3·s–1). For this SI leakage rate
to be a mass flow, the pressure and
temperature must be at standard values of
101 kPa (760 torr) and 0 °C (32 °F).
Table 11 gives factors for conversion of
TABLE 11. Mass flow conversion factors
for leakage rate.
To Convert from
To
Pascal cubic meter per std cm3·s–1
second (Pa·m3·s–1)
mol·s–1
torr·L·s–1
mb·L·s–1
std ft3·h–1
Standard cubic
Pa·m3·s–1
centimeter per
mol·s–1
second (std cm3·s–1) torr·L·s–1
mb·L·s–1
std ft3·h–1
Mole per second
Pa·m3·s–1
(mol·s–1)
std cm3·s–1
torr·L·s–1
mb·L·s–1
std ft3·h–1
Torr liter per second
Pa·m3·s–1
(torr·L·s–1)
std cm3·s–1
mol·s–1
mb·L·s–1
std ft3·h–1
Millibar liter per
Pa·m3·s–1
second (mb·L·s–1)
std cm3·s–1
mol·s–1
torr·L·s–1
std ft3·h–1
Standard cubic foot per Pa·m3·s–1
hour (std ft3·h–1)
std cm3·s–1
mol·s–1
torr·L·s–1
mb·L·s–1
Multiply by
9.87 (≅ 10)
4.40 × 10–4
7.50
1.00 × 101
1.25
1.01 × 10–1
4.46 × 10–5
7.60 × 10–1
1.01
1.27 × 10–1
2.27 × 103
2.24 × 104
1.70 × 104
2.27 × 105
2.85 × 103
1.33 × 10–1
1.32
5.87 × 10–5
1.33
1.67 × 10–1
1.00 × 10–1
9.87 × 10–1
2.27 × 104
7.50 × 10–1
1.26 × 10–1
0.80
7.87
3.51 × 10–4
5.99
7.94
leakage rates in various common units,
past and present. Table 12 provides
leakage rate comparisons that permit a
better understanding of the quantities
involved, when leakage rates are specified.
Leakage is not simply the volume of air
entering the vacuum chamber. Instead,
the critical factor is the number of gaseous
molecules entering the vacuum system.
This number of molecules, in turn,
depends on the external pressure,
temperature and the volume of gas at this
pressure that leaks into the vacuum
system. The leakage rate is expressed in
terms of the product of this pressure
difference multiplied by the gas volume
passing through the leak, per unit of time.
Thus, the leakage rate is directly
proportional to the number of molecules
leaking into the vacuum system per unit
of time.
The molecular unit of mass flow used
for gas by the National Institute of
Standards and Technology is mole per
second (mol·s–1), a mass flow unit
measured at standard atmospheric
pressure and standard temperature of 0 °C
(32 °F). A common unit of gas is the
standard cubic meter (std m3). This unit is
equivalent to one million units given as
atmospheric cubic centimeter (atm cm3).
Both units indicate the quantity of gas
(air) contained in a unit volume at
average sea level atmospheric pressure at a
temperature of 0 °C (32 °F). The average
atmospheric pressure at sea level is
101.3 kPa (760 mm Hg or 760 torr). The
SI unit of pressure, the pascal (Pa), is
equivalent to newton per square meter
(N·m–2).
Non-SI Units Used Earlier
for Measurement of
Leakage
Various units have been used for
measurement of leakage, including many
related to English units commonly used in
engineering in the United States.
Justification for prior use of this diversity
of units lies in the relative ease with
which these common units can be
adapted for many practical engineering
problems. For example, suppose that an
operator has a gas cylinder with a pressure
gage calibrated in units of pound-force per
square inch (lbf·in.–2). With daily gage
readings, it is convenient for the operator
to express leakage as the gage pressure
change multiplied by cylinder volume,
divided by the leakage time period (days).
This simple calculation results in leakage
rate measurement in units of lbf·in.–2 ft3
per day. This leakage rate has dimensions
of (pressure) × (volume) ÷ (time). To have
expressed the leakage merely as the
Introduction to Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
29
volume of gas lost is insufficient because
the volume of gas that leaves daily at high
cylinder pressure will be considerably
larger than the volume leaking to the
atmosphere each day when the internal
pressure of the cylinder is lower. Many
combinations of units for pressure,
volume and time are possible. The SI
volumetric leakage rate unit pascal cubic
meter per second (Pa·m3·s–1) is used in
this book.
Units for Leakage Rates of
Vacuum Systems
Suppose that leakage of air into a vacuum
system has an undesired effect on the
pressure within the vacuum system. The
operator of the vacuum system can read
absolute pressures in pascal or torr from
gages permanently installed in the system.
(The pressure unit known as a torr is
defined as 1/760th of a standard
atmosphere and differs only by one part
in seven million from the well known
barometric pressure unit of millimeter
mercury.) In the past, the leakage rate in
vacuum systems was measured in torr liter
per second. If the volume of the vacuum
chamber had been measured in cubic
meter, the operator might find it easier to
measure leakage rate in units of pascal
cubic meter per day or per second.
Leakage is not simply the volume of air
entering the vacuum chamber. Instead,
the critical factor is the number of gaseous
molecules entering the vacuum system.
This number of molecules, in turn,
depends on the external pressure,
temperature and the volume of gas at this
pressure that leaks into the vacuum
system. The leakage rate is expressed in
terms of the product of this pressure
difference multiplied by the gas volume
passing through the leak, per unit of time.
Thus, the leakage rate is directly
proportional to the number of molecules
leaking into the vacuum system per unit
of time.
30
Leak Testing
TABLE 12. Leakage rates expressed in
various units
Pa·m3·s–1
1
10–1
10–2
10–3
10
10–5
10–6
10–7
10–8
10–9
10–10
std cm3·s–1
10
1
10–1
10–2
10–3
10–4
10–5
10–6
10–7
10–8
10–9
mol·s–1
4.40
4.40
4.40
4.40
4.40
4.40
4.40
4.40
4.40
4.40
4.40
×
×
×
×
×
×
×
×
×
×
×
10–4
10–5
10–6
10–7
10–8
10–9
10–10
10–11
10–12
10–13
10–14
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
References
1. Nondestructive Testing Handbook,
second edition: Vol. 10, Nondestructive
Testing Overview. Columbus, OH:
American Society for Nondestructive
Testing (1996).
2. Wenk, S.A. and R.C. McMaster.
Choosing NDT: Applications, Costs and
Benefits of Nondestructive Testing in Your
Quality Assurance Program. Columbus,
OH: American Society for
Nondestructive Testing (1987).
3. Nondestructive Testing Methods.
TO33B-1-1 (NAVAIR 01-1A-16)
TM43-0103. Washington, DC:
Department of Defense (June 1984).
4. Nondestructive Testing Handbook,
second edition: Vol. 1, Leak Testing.
Columbus, OH: American Society for
Nondestructive Testing (1982).
5. Marr, J.W. Leakage Testing Handbook.
Report No. CR-952. College Park, MD:
National Aeronautics and Space
Administration, Scientific and
Technical Information Facility (1968).
6. E 432-91, Standard Guide for Selection of
a Leak Testing Method. West
Conshohocken, PA: American Society
of Testing and Materials (1996).
7. Waterstrat, C. “The Need to Train Leak
Testing Personnel.” Materials
Evaluation. Vol. 47, No. 11. Columbus,
OH: American Society for
Nondestructive Testing (November
1989): p 1263-1265.
8. Recommended Practice No. SNT-TC-1A.
Columbus, OH: American Society for
Nondestructive Testing (1996).
9. Nerken, A. “History of Leak Testing.”
Materials Evaluation. Vol. 47, No. 11.
Columbus, OH: American Society for
Nondestructive Testing (November
1989): p 1268-1272.
10. Prout, H.G. A Life of George
Westinghouse. New York, NY: American
Society of Mechanical Engineers
(1921).
11. IEEE/ASTM SI 10-1997, Standard for Use
of the International System of Units (SI):
The Modernized Metric System.
Philadelphia, PA: American Society for
Testing and Materials (1996).
12. Jakuba, S. Metric (SI) in Everyday Science
and Engineering. Warrendale, PA:
Society of Automotive Engineers
(1993).
13. Taylor, B.N. Guide for the Use of the
International System of Units (SI). NIST
Special Publication 811, 1995 edition.
Washington, DC: United States
Government Printing Office (1995).
Introduction to Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
31
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
C
2
H A P T E R
Tracer Gases in Leak
Testing1
Charles N. Sherlock, Willis, Texas
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
PART 1. Introduction to Properties of Tracer
Gases for Leak Testing
Fluid Media Used in Leak
Testing
Leak testing can be divided into three
categories: (1) leak detection, (2) leak
location and (3) leakage measurement.
Each involves a fluid leak tracer and some
means for establishing a pressure
differential or other means to make fluid
flow through the leak or leaks. Possible
fluid probing media include gases, vapors,
liquids or combinations of these. Selection
of the desired fluid probing medium for
leak testing depends on operator or
engineering judgment and involves
factors such as:
1. type and size of test object or system
to be tested;
2. typical operating conditions of test
object or system;
3. environmental conditions during leak
testing;
4. hazards associated with the probing
medium and the pressure involved in
testing;
5. leak testing instrumentation and its
response to the probing medium; and
6. leakage rates that must be detected
and the accuracy with which
measurements must be made.
Where high sensitivity to leakage must
be attained, gases and vapors are generally
preferred to liquid media. The present
discussion is devoted specifically to
gaseous tracers used in leak testing.
Special gaseous tracers are discussed
elsewhere in this volume. Liquid probing
media are used for leak testing in many
applications.
Volumes Occupied by
Gases and Liquids
The volume of any substance is the space
occupied by that substance. For gases, the
volume of a sample of gas is the same as
the volume of the container within which
the gas is held. The volume occupied by
liquids or by solids does not change very
much with a change in pressure or
temperature. Therefore, to describe the
amount of a solid or of a liquid, it is
usually sufficient to specify only the
volume of the sample. However, this
cannot be done with gases. For example,
34
Leak Testing
1 m3 of gaseous helium at a certain
temperature and pressure will have a
different gas density and mass than would
1 m3 of gaseous helium at different
temperature and pressure conditions. To
determine the quantity of a given volume
of gas, it is necessary to know its pressure
and temperature. When liquids are mixed
together, the total volume is roughly
equal to the sums of the original volumes.
However, this is not necessarily true for
mixtures of gases. Gases can mix in any
proportions and still fill the volumes
within which they are mixed.
Pressures Exerted by Gases
or Liquids
Fluid pressure is defined as a force per unit
area. In liquids and gases, the pressure at a
given point is the same in all directions. In
general, for all gases and liquids, the
greater the depth of immersion, the
greater the internal pressure. These effects
can be illustrated by considering a
swimmer under water. At a given depth,
the pressure exerted on the body is the
same no matter how the swimmer turns.
This is due to the pull of gravity on the
water above. The body is subject to
pressure because it must support the
weight of water above the swimmer.
The earth is surrounded by a blanket of
air several hundred kilometers thick.
People live at the bottom of this ocean of
air, which exerts atmospheric pressure.
The force per unit area exerted on the
earth’s surface is equal to the weight of
the column of air above it, 100 kPa
(14.7 lbf·in.–2). This pressure also
corresponds to the weight of a column of
mercury 760 mm high, or 760 torr. The
mercury barometer balances the weight of
its column of mercy against the weight
per unit area of the earth’s atmosphere. At
sea level, the pressure is typically near
100 kPa (14.7 lbf·in.–2). The pressure is
reduced as the altitude increases, so the
barometer can also be used as an
altimeter. The atmospheric pressure also
changes from day to day as cold, dense air
masses are replaced by less dense warm air
masses and vice versa. Thus, care must be
taken to exclude the effects of local
changes in atmospheric pressure from leak
testing measurements or to correct for
their effects.
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
Pressures can be measured in
atmospheres (atm) with respect to zero
pressure (absolute pressures) or normal
atmospheric pressure (gage pressures). In
general, gas pressure is a measure of the
work done to compress gas into a unit
volume. The change in energy W stored
in gas under pressure within a container is
related to the product P of its pressure and
its volume V, as in Eq. 1:
(1)
W
= PV
where P is absolute pressure of gas
(pascal), V is volume of gas (cubic meter)
and W is stored energy (joule).
Boyle’s Law Relating
Pressure and Volume of
Gases at Constant
Temperature
A characteristic property of gases is that
they are easily compressed. This behavior
is described by Boyle’s law (1662), which
states that, at constant temperature, a
fixed mass of gas occupies a volume
inversely proportional to the pressure
exerted on it. If the pressure is doubled,
the volume becomes half as large (Fig. 1).
Boyle’s law is expressed by Eq. 2:
(2)
Pi Vi
=
(3)
Vi
Vf
=
Ti
Tf
where the subscripts i and f denote the
initial and the final conditions,
respectively. In Eq. 3, the temperature T
must always be expressed in units of
absolute temperature (kelvin or degree
rankine).
Variations of temperature of contained
gases during leak testing could lead to
erroneous interpretations of leak testing
data if the effects of Charles’s law were
ignored. Thus, it is desirable to make leak
tests during periods of reasonably
constant temperature, if possible, and to
correct for test temperature variations
during data analysis to ensure valid
interpretations and measurements of
leakage.
Dalton’s Law of Partial
Pressures of Mixed Gases
The behavior observed when two or more
gases are placed within the same
container is summarized in Dalton’s law
FIGURE 1. Boyle’s law experiment showing volume decrease
of gas when pressure increases, at constant temperature.
Pf Vf
In Eq. 2, the subscripts i and f denote
the initial and final conditions,
respectively, of the fixed quantity or mass
of gas.
Charles’ Law Relating
Temperature and Volume
of Gases under Constant
Pressure
Like most substances, gases increase in
volume when their temperature is raised.
This increase in volume with increasing
temperature can be observed
experimentally with the arrangement
sketched in Fig. 2. If the force on top of
the piston is constant, the gas sample
remains at constant pressure P. If the gas
is heated, the piston moves out and the
volume V of gas beneath it increases. This
behavior is expressed by Charles’ law
(1787), which states that, at constant
pressure, the volume occupied by a fixed
mass of gas is directly proportional to the
absolute (kelvin) temperature of the gas.
Mathematically, Charles’s law is expressed
by Eq. 3:
Force = F
Force = 2F
Volume = 1 m3
1m
Volume = 0.5 m3
0.5 m
FIGURE 2. Charles’ law experiment showing volume increase
with temperature, in gas at constant pressure.
Force = F
Force = F
Volume = 1 m3
1m
Volume = 0.5 m3
Temperature = 400 K
Temperature = 800 K
0.5 m
Tracer Gases in Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
35
of partial pressures (1801), which states
that the total pressure exerted by a
mixture of gases is equal to the sum of the
partial pressures of the various gases. The
partial pressure of a gas in a mixture is
defined as the pressure the specific gas
would exert if it were alone in the
container. The meaning of Dalton’s law is
indicated by the sketch of Fig. 3. One
cubic meter (1.0 m3 or 35 ft3) of nitrogen
at a pressure of 50 kPa (7.25 lbf·in–2) and
1.0 m3 (35 ft3) of oxygen at a pressure of
70 kPa (10.15 lbf·in–2) would exert a total
pressure of 120 kPa (17.40 lbf·in–2) if the
two gases were mixed and contained
within a volume of 1.0 m3 (35 ft3). For the
general case, Dalton’s law can be
expressed by Eq. 4:
(4)
Ptotal
=
P1 + P2 + P3 + … Pn
FIGURE 3. Dalton’s law experiment showing total pressure to
equal sum of partial pressures of mixed gases injected into a
fixed volume: (a) oxygen; (b) nitrogen; (c) combined
pressure of same quantitites of nitrogen and oxygen
combined.
P = 50 kPa
(7 lbf·in.–2)
(a)
Oxygen
Volume = 1 m3
P = 70 kPa
(10 lbf·in.–2)
(b)
Nitrogen
where the numerical subscripts indicate
the partial pressures due to each gas
constituent.
Volume = 1 m3
P = 120 kPa
(17 lbf·in.–2)
Avogadro’s Principle
Describing Number of Gas
Molecules in a Volume
Amedeo Avogadro in 1811 was the first to
propose the principle now known as
Avogadro’s principle. It states that equal
volumes of gases at the same temperature
and pressure contain equal numbers of
gas molecules. Through modern
techniques it has been possible to make
the following observation concerning the
average number of gas molecules in one
mole of gas. A mole is the amount of gas
whose weight in gram equals its atomic
mass. Avogadro’s number of 6 × 1023
molecules (a mole) is the number of gas
molecules that would occupy a volume of
22.4 L (0.79 ft3) at standard temperature
and pressure. Standard temperature is
designated at 0 °C (32 °F), the freezing
point of water; standard pressure is
defined as 100 kPa (1 atm). This standard
pressure was originally based on the
atmospheric pressure that will support a
column of mercury 760 mm in height,
which corresponds to the mean
atmospheric pressure at sea level.
According to Avogadro’s principle, the
volume that a gas sample occupies at
standard temperature and pressure is
directly proportional to the number of gas
molecules within that gas sample.
(c)
Nitrogen and oxygen
Volume = 1 m3
General Gas Law
Applicable to All Ideal
Gases and Mixtures of
Ideal Gases
Boyle’s law, Charles’ law and Avogadro’s
principle can be combined to give a
general relationship between volume V,
pressure P, temperature T and the number
of moles of gas m in a gas sample. The
general gas law can be applied without
the necessity of maintaining one of these
variables constant. Boyle’s law states that
the volume occupied by a gas is inversely
proportional to the gas pressure. Charles’
law states that the gas volume is directly
proportional to the gas temperature.
Avogadro’s principle states that the
volume is directly proportional to the
total number of gas molecules contained
in that volume (regardless of the species
of the individual molecules). These
relationships are summarized in Eqs. 5
through 8, in which the symbol ≅ means
“is proportional to”: Boyle’s law,
(5)
V
≅
1
P
with constant T and m; Charles’ law,
36
Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
LT.02 LAYOUT 11/8/04 2:13 PM Page 37
(6)
V
≅
T
with constant P and m; Avogadro’s
principle,
(7)
V
≅
m
with constant T and P; and a General
relationship,
(8)
V
mT
≅
P
without restriction.
The general relationship of Eq. 8
combines the individual relationships of
Eq. 5, 6 and 7. This can be seen by
imagining that any two of the variables,
such as T and m, are constant and noting
the relation of the other two variables.
The general ideal gas law (applicable to
all ideal gases) can be written in the form
of Eq. 9:
(9)
=
PV
m RT
Here, R (in J·mol–1·K–1) is the universal
gas constant found from known values of
P, V, m and T by Avogadro’s principle, by
use of EQ. 10:
(10)
R
=
PV
mT
=
8.314
The individual gas constant Ri
(J·kg–1·K–1) is obtained by dividing the
universal gas constant R (joule) by the
molecular mass M (kilogram) of the
specific gas involved, by use of Eq. 11:
(11)
Ri
=
R
M
=
The numerical value of the individual gas
constant for several common tracer gases
is given in Table 1. The behavior of real
gases conforms closely to the Ideal gas law
of Eq. 9 under a wide range of conditions.
It begins to deviate from this ideal gas law
only as gas densities become much higher
than those usually used in leak testing.
However, the behavior of vapors,
including several types of vapors used in
leak testing, can deviate significantly from
the relation of the Ideal gas law. Thus,
care is required in computing leakage
rates by the ideal gas law relationship
when the pressurizing gas or leak tracer is
a vapor or contains a large proportion of
vapor constituent. (A vapor is the gaseous
form of any substance that usually exists
in the form of a liquid or a solid, such as
water vapor. A pure liquid in equilibrium
with its own vapor will have two phases,
liquid and vapor, which coexist at a
specific partial pressure known as the
vapor pressure. Because condensation or
evaporation can occur, vapor molecules
can enter or leave the gaseous phase. This
changes the number of molecules of that
vapor species that will be present within a
particular gas volume.) These vapor effects
are not included in the general gas law
relationship of Eqs. 9 to 11.
PV
mMT
Graham’s Law for Diffusion
of Gases
A gas expands to occupy the volume
within which it is contained. If a bottle of
ammonia is opened at one end of a room,
it is soon detected by odor at the other
end of the room. This spreading of a gas
constituent through other gaseous
TABLE 1. Physical properties of typical gases and vapors at 15 °C (59 °F).
Gas
Air
Ammonia
Argon
Carbon dioxide
Dichlorodifluoromethane
Helium
Hydrochloric acid
Hydrogen
Krypton
Methane
Neon
Nitrogen
Nitrous oxide
Oxygen
Sulfur dioxide
Water vapor
Chemical
Symbols
NH3
Ar
CO2
CCl2F2
He
HCI
H2
Kr
CH4
Ne
N2
N2O
O2
SO2
H2O
Molecular
Molecular
Weight
Diameter (pm)
29.00
17.03
39.94
44.01
120.93
4.00
36.50
2.02
83.80
16.04
20.18
28.01
44.00
31.99
64.00
18.02
297.0
288.0
334.0
190.0
240.0
315.0
298.0
460.0
Viscosity
(µPa·s)
Gas Constant,
(J·kg–1·K–1)
18.0
9.7
22.0
14.5
12.7
19.2
14.0
8.6
24.6
10.7
31.0
17.3
14.3
19.9
12.3
9.3
287
488
208
189
68.8
2079
228
4116
9.92
518
412
297
189
260
130
461
Tracer Gases in Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
37
constituents within a volume is called
diffusion. Under fixed conditions, it is
found that lighter gases diffuse more
rapidly than the heavier gases. Graham’s
law of diffusion states, The rates of
diffusion of different gases are inversely
proportional to their individual molecular
masses. Graham’s law can be written
mathematically in the form of Eq. 12:
(12)
D1
D2
=
M1
M2
where D1 and D2 are the rates of diffusion
of gases 1 and 2 and where M1 and M2 are
the respective molecular masses of these
two different gases.
A leak testing tracer gas with low
diffusivity provides an advantage in
detector probe leak detection techniques
because the concentration of the tracer
gas builds up at the leak exit. This allows
detection by a probe to locate the site of a
leak. With low diffusivity, the tracer gas
does not leave the leak location rapidly.
A tracer gas of high diffusivity is
needed for internal pressurization where it
is necessary to fill cul-de-sacs or blind
passageways within a reasonable soak
time before testing. A low diffusion rate
would not allow a tracer gas to traverse a
tortuous leak passage, thus making leak
detection by tracer gas an unreliable
procedure. Table 2 lists the diffusivities of
typical tracer gases in air at standard
TABLE 2. Diffusivities of tracer gases in air at standard temperature of 0 °C (32 °F) and
standard pressure of 100 kPa (760 torr). (Diffusion coefficient values are calculated from
an empirical equation, after Slattery.2)
Molecular
Mass
38
Leak Testing
Difffusion
Coefficient
Gas
Formula
(g·mol–1)
mm2·s–1
(ft2·h–1)
Acetylene
Ammonia
Argon
Benzene
Butane
Carbon dioxide
Carbon disulfide
Carbon monoxide
Carbon tetrachloride
Dichloromethane
Ethane
Ethyl alcohol
Ethylene
Refrigerant–11
Refrigerant–12
Refrigerant–21
Refrigerant–22
Refrigerant–112
Refrigerant–114
Refrigerant-134a
Helium
Hydrogen
Hydrogen sulfide
Krypton
Methane
Neon
Nitric oxide
Nitrogen
Nitrous oxide
Oxygen
Propane
Sulfur dioxide
Sulfur hexafluoride
Water
Xenon
C2H2
NH3
Ar
C6H6
C4H10
CO2
CS2
CO
CCl4
CH2Cl2
C2H6
C2H5OH
C2H4
CCl3F
CCl2F2
CHCl2F
CHClF2
CCl2F–CCl2F
CClF2–CClF2
C2H2F4
He
H2
H2S
Kr
CH4
Ne
NO
N2
N2O
O2
C3H8
SO2
SF6
H2O
Xe
26.0
17.0
39.9
78.1
58.1
44.0
76.1
28.0
154.0
84.93
30.1
46.1
28.0
137.0
121.0
103.0
86.5
204.0
171.0
102.0
4.0
2.0
24.1
83.8
16.0
20.2
30.0
28.0
44.0
32.0
44.1
64.1
146.0
18.0
131.0
14.2
17.0
14.7
7.7
8.5
13.4
9.3
17.3
7.2
7.4
12.6
9.8
13.4
7.7
8.3
8.5
9.5
6.5
7.2
7.2
69.7
67.1
13.7
13.2
18.6
28.4
18.1
17.5
13.4
17.5
10.0
10.8
7.3
21.9
10.8
(0.55)
(0.66)
(0.61)
(0.30)
(0.33)
(0.52)
(0.36)
(0.67)
(0.28)
(0.29)
(0.49)
(0.38)
(0.52)
(0.30)
(0.32)
(0.33)
(0.37)
(0.25)
(0.28)
(0.28)
(2.70)
(2.60)
(0.53)
(0.51)
(0.72)
(1.10)
(0.70)
(0.68)
(0.52)
(0.68)
(0.39)
(0.42)
(0.28)
(0.85)
(0.42)
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
conditions of 100 kPa (1 atm) pressure
and a temperature of 0 °C (32 °F). The
diffusion coefficient values in Table 2 are
calculated and converted from an
empirical equation.2
Brownian Motion of Gases
One aspect of gaseous behavior that gives
the strongest clue to the nature of gases is
the phenomenon known as Brownian
motion. This motion, first observed by the
Scottish botanist Robert Brown in 1827, is
the irregular motion of extremely minute
particles suspended in a fluid. Brownian
motion can be observed by focusing a
microscope on a particle of illuminated
cigarette smoke in a glass tube. The
particle does not settle to the bottom of
the container but continues to move
randomly in all directions. The smaller
the suspended particle under observation,
the higher the temperature of the fluid,
the more vigorous is the particle’s
movement. The existence of Brownian
motion suggests that the molecules of
gaseous matter are constantly moving. A
visible small particle seems to be jostled
by its neighboring invisible particles. The
motion of the visible smoke particle thus
indirectly reflects the motions of the
smaller invisible particles of matter. This
provides powerful support for the idea
that gaseous or fluid matter consists of
extremely small particles or molecules in
constant motion. The theory of moving
molecules of gases is the kinetic molecular
theory of matter. Its basic postulates are
these.
1. The molecules of gaseous matter are in
motion.
2. Heat causes this molecular motion.
The kinetic theory of gases can be used to
explain many of the properties and
characteristics of tracer gases used in leak
testing.
Assumptions Underlying
the Kinetic Theory of Ideal
Gases
The kinetic theory of gases applies only to
ideal or perfect gases that behave in
accordance with the following
assumptions.
1. Gases consist of tiny molecules so
small and so far apart that the actual
volume of the gas molecules is
negligible compared to the empty
space between them.
2. There are no attractive forces between
gaseous molecules.
3. The molecules of gases travel in
random straight-line motion and
collide elastically with each other and
with the walls of their container.
4. In any collection of gas molecules,
individual molecules have different
speeds. However, their average speed
(including many molecules over a
significant period of time) is
dependent on the absolute
temperature (kelvin or rankine
degrees). The higher the gas
temperature, the higher the average
molecular speed.
Kinetic Theory Explanations of
Gaseous Pressure, Volume and
Temperature
The kinetic theory of gases postulates that
a gas consists mostly of empty space in
which billions of tiny points representing
molecules are moving randomly. The
molecular particles collide with each other
and with the walls of the container. The
volume of a gas sample is the volume of
its container. Pressure is exerted by gases
because the molecules collide with the
walls of the container. Each collision
produces a tiny push or impulse as the
molecule rebounds from the wall. The
sum of all of these molecular pushes or
force impulses of impact constitute the
pressure of the gas on its containment
walls. The temperature of a gas is a
measure of the average speed or kinetic
energy of the particles.
Kinetic Theory Explanations of the
Gas Laws
The kinetic molecular theory of gases can
be used to explain the observed behavior
of gases as described by the gas laws.
Boyle’s Law. The pressure exerted by a gas
at a given temperature depends only on
the number of impacts of gas molecules
with the walls of the container. If the
volume is reduced as sketched in Fig. 4,
FIGURE 4. Example of Boyle’s law. Doubling of gas pressure
concentrates gas molecules and doubles number of
molecular impacts per unit area on chamber walls and piston
in given time period.
Force = F
Force = 2F
Volume = 1 m3
Volume = 0.5 m3
Tracer Gases in Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
39
the molecules are more confined. This
increases the frequency of molecular
collisions with the walls. These more
numerous impacts are observed as a
greater pressure.
Charles’ Law. If the temperature of a gas
rises, the average molecular energy and so
the average speed of the gas molecules
rises. As the molecules move more
energetically, they collide with the walls
of the container more frequently and with
greater momenta, thus producing greater
pressure. (Force is equal to the time rate of
change of momentum and pressure is the
force per unit area.) As shown in Fig. 5, if
the temperature is raised, the balloon
responds to the increased pressure by
stretching and expanding its diameter.
Dalton’s Law. According to the kinetic
theory of ideal gases, there are no
attractive forces between the molecules of
gases. On the average, the molecules of
each constituent of a gaseous mixture will
strike the walls of their container the
same number of times per second and
strike with the same impact forces as they
would if there were no other gases
FIGURE 5. Example of Charles’ law. Raising
gas temperature increases molecular
velocities and increases gas pressure on
container wall. Under constant atmospheric
pressure, impacts by higher velocity
molecules cause increase in gas volume
within elastic balloon.
Heated
balloon
Cooled
balloon
constituents present (see Fig. 6.)
Therefore, the partial pressure of a gaseous
constituent in a gas mixture is not
changed by the presence of other gases in
the container. The total pressure exerted
on the walls of the container (or on the
diaphragm of a pressure measuring gage)
is equal to the sum of the partial pressures
exerted by the individual constituents of
the gaseous mixture.
Determining
Concentration of Tracer
Gas in Gas Mixtures from
Partial Pressures
In many leakage measurements, it is
desirable or necessary to dilute the tracer
gas being used for leak testing. Use of
diluted tracer gas might be dictated by
practical considerations such as:
1. high expense of pure tracer gas filling
large volumes or attaining high
pressures;
2. attainment of a more nearly linear or
more stable instrument response at a
lower concentration of tracer gas;
3. pure tracer gas providing a much
higher leakage sensitivity than needed;
4. danger of fire or explosion with a
flammable tracer gas (in some cases, a
dilute gas mixture lowers the danger
of explosions); and
5. inability to completely evacuate the
test object or test system before filling
with tracer gas. As a result, residual gas
dilutes the tracer gas added during
pressurizing.
Concentration of the tracer gas in a
test system containing mixed gases
depends on the partial pressure of the
tracer gas. Dalton’s law (Eq. 4) shows the
contributions of each gaseous constituent
to the total gas pressure. The fractional
concentration of the tracer gas T is given
by the term NT in Eq. 13:
FIGURE 6. Example of Dalton’s law: (a) oxygen; (b) nitrogen; (c) nitrogen and oxygen
combined. Partial pressure of each gaseous constituent is not changed by presence of other
gases in the same container. Pressure is exerted on container walls by impacts of individual
molecules of all gas species.
(a)
N2
O2
P=5
40
Leak Testing
(c)
(b)
N2 and O2
P=7
P = 12
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
(13) NT
PT
=
P total
PT
=
P1 + P2 + P3 + … + Pn
This fractional concentration is given
by Eq. 13 in terms of the number of
molecules of tracer gas as a fractional part
of the total number of molecules in a
gaseous mixture. The percentage
concentration by mass would depend also
on the molecular masses of each gaseous
constituent. The partial pressure PT of
tracer gas required to provide a specific
percentage of tracer gas molecules,
percent T, is given in terms of total
system pressure Ptotal by Eq. 14:
(14)
PT
=
%T × Ptotal
100
Partial Pressures of Gaseous
Constituents of Earth’s
Atmosphere
that in the atmosphere at sea level.
Helium is present in the earth’s
atmosphere in the proportion of 5 µL·L–1.
With mass spectrometer types of helium
leak detectors, even this small proportion
of helium can be readily sensed.
Mean Free Paths of Gas
Molecules
The mean free path is the average
distance a gas molecule travels between
successive collisions with other molecules
in the gas or vapor state. The mean free
path is important in leak testing because
it determines the type of gas flow that will
occur through leaks or other passageways
traversed by tracer or pressurizing gases.
The mean free path can be calculated
from the pressure, temperature and
molecular properties by means of Eq. 15:
(15)
The composition of atmospheric air is
78 percent nitrogen, 21 percent oxygen,
0.9 percent argon and about 0.1 percent
of other gases and vapors (including water
vapor, whose concentration varies with
the temperature and relative humidity of
the atmosphere). The partial pressures of
atmospheric constituents at sea level,
where the total pressure of 100 kPa
(1 atm) is equal to that of 760 torr, are
given in Table 3.
The partial pressure in kilopascal is
about the same as the percentage of each
constituent gas, at standard atmospheric
pressure. The partial pressures of
atmospheric constituents at an altitude of
3600 m (12 000 ft), where the total
pressure is equal to 64.4 kPa (9.3 lbf·in.–2
absolute), are given in Table 2.
The percentage composition of
atmospheric air changes very little until
very high altitudes are reached. When test
systems are pressurized with air pumped
from the atmosphere, the percentage
composition is also not changed from
λ
=
116.4
n
P
T
M
where λ is mean free path (meter) under
static pressure; n is gas viscosity (pascalsecond); P is absolute pressure of gas
(pascal); T is absolute gas temperature,
(kelvin); and M is molecular mass of gas,
(g-mol–1).
Table 4 lists the mean free paths of
common gases at 20 °C (68 °F) and
1.0 mPa (7.6 µtorr).
Simple Approximation Formula
for Mean Free Path of Common
Gases
An easily remembered relation for
approximating the mean free paths of
common gaseous molecules is presented
in Eq. 16:
(16)
λ
=
NF
P
where λ is mean free path length (meter),
P is gas pressure in pascal and NF is a
numerical factor (meter-pascal) given in
Table 5. An NF value of 6.8 × 10–3 permits
TABLE 3. Partial pressures of atmospheric constituents at sea level, 100 kPa (1 atm).
Gas
Oxygen (O2)
Nitrogen (N2)
Argon (Ar)
Other
Total air
Percent
21.0
78.0
0.9
0.1
100.0
kPa at
sea levela
(torr at
sea level)a
kPa at
3.6 km
(torr at
12 000 ft)b
21.28
79.03
0.91
0.10
101.325
(159.6)
(592.8)
(6.84)
(0.76)
(760.00)
13.52
50.22
0.58
0.06
64.4
(101.4)
(376.65)
(4.35)
(0.45)
(483.0)
a. Atmospheric pressure at sea level = 101.325 kPa (1 atm).
b. Percentage × atmospheric pressure of 64.4 kPa.
Tracer Gases in Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
41
LT.02 LAYOUT 11/8/04 2:14 PM Page 42
TABLE 4. Mean free path lengths of
various atmospheric gases at 20 °C
(68 °F) and at absolute pressure of
1.0 mPa (7.6 µtorr).
Mean Free Path
___________________
m
(in.)
Gas
Air
6.8
Argon (Ar)
7.2
Carbon dioxide (CO2) 4.5
Hydrogen (H2)
12.5
Water (H2O)
4.2
Helium (He)
19.6
Nitrogen (N2)
6.7
Neon (Ne)
14.0
Oxygen (O2)
7.2
(268)
(284)
(177)
(492)
(165)
(771)
(264)
(551)
(284)
use of Eq. 16 for estimating the mean free
path lengths for air, argon, nitrogen and
oxygen molecules. Table 5 lists the
numerical factors used in the numerator
of Eq. 16 for several other common gases.
However, it is seldom necessary to know
mean free path lengths to precisions
better than one order of magnitude. For
example, the molecular mean free path at
20 °C (68 °F) at atmospheric pressure is of
the order of 30 to 300 nm. At a pressure
of 1 Pa (1.5 × 10–4 lbf·in.–2), the mean free
path is in the range from 3 to 30 mm
(0.12 to 1.2 in.).
Relation of Mean Free Path
Lengths to Viscosity and
Molecular Mass of Gas
The ratio of mean free path lengths for
two different gases at the same
temperature and pressure are given by
Eq. 17:
(17)
λ1
λ2
=
n1
n2
M2
M1
In Eq. 17, n indicates the gas viscosity
and M indicates its molecular mass. The
subscript 1 indicates the first gas and
subscript 2 indicates the second gas. (This
relationship is derived from Eq. 15 when
T and P are held constant.) For a leak or
an orifice across which there is a sizable
pressure differential, the mean free path
length within the orifice is typically
estimated from the average pressure in the
orifice (the mean value of inlet and outlet
pressures).
Effective Viscosity of Mixtures of
Gases
In a mixture of various species of gases,
the effective viscosity nmixture is assumed
to be proportional to the sum of the
products of viscosity and concentrations
for each individual gaseous constituent, as
indicated by Eq. 18:
TABLE 5. Physical properties of common gases used in leak testing.
Gas
Formula
Airf
Argon
Carbon dioxide
Refrigerant-12
Helium
Hydrogen
Krypton
Neon
Nitrogen
Oxygen
Sulfur hexafluoride
Water Vaporg
Xenon
a.
b.
c.
d.
e.
f.
g.
Mixture
Ar
CO2
CCl2F2
He
H2
Kr
Ne
N2
O2
SF2
H2O
Xe
Densitya
Molecular
at
Mass
100 kPa
(g·mol–1) (g·L–1)
29.0
40
44
121
4.0
2.0
84
20
28
32
146
18
131
1.21
1.79
1.97
5.25
0.179
0.090
3.74
0.90
1.25
1.43
6.60
0.83
5.89
Numerical
Factorb for
Dynamic
Mean Free
Viscosityc
Path
at 20 °C (68 °F)
(m·Pa)
(µPa·s)
6.8 × 10–3
7.2 × 10–3
4.5 × 10–3
19.6 × 10–3
12.5 × 10–3
5.36 × 10–6
14.0 × 10–3
6.7 × 10–3
7.2 × 10–3
2.5 × 10–3
4.2 × 10–3
3.8 × 10–3
18
22
15
13
19
9
25
31
18
20
15
9
22
Diffusivityd
in Air at
0 °C (32 °F)
and 101 kPa
(m2·s–1)
13.9 × 10–6
15.8 × 10–6
63.4 × 10–6
17.8 × 10–6
23.9 × 10–6
Thermal
Conductivitye
at 20 °C (68 °F)
(W·m–1·K–1)
26.2
17.9
16.0
9.8
149.0
183.0
9.4
48.0
25.6
26.2
13.0
18.7
5.5
Density in oz·ft–3 = g·L–1 = mg·cm–3 at 20 °C (68 °F) and 100 kPa (1 atm).
Numerical factor for calculating mean free path using Eq. 16. Mean free path in meters at 20 °C (68 °F).
Independent of pressure under conditions for viscous flow.
Diffusivity in m2·s–1 in air at 0 °C (32 °F) and 101 kPa (1 atm).
Thermal conductivity in W·m–1·K–1 at 20 °C (68 °F). Thermal conductivity is independent of pressure under conditions for viscous flow.
N2, 78 percent; O2, 21 percent; argon, 0.9 percent; others, 0.1 percent.
Vapor pressure of H2O at 20 °C (68 °F) is 2.3 kPa (17.5 torr).
42
Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
(18) n mixture
= N1 n1 + N2 n 2
+ … + Nk nk
where n1 is viscosity of the first gaseous
constituent and N1 is fractional
concentration of the first gaseous
constituent, as defined by Eq. 14.
Molecular Masses of Gases
and Vapors
Any combination of atoms in a chemical
compound is called a molecule. The
molecular mass equals the total number
of nucleons in the atoms forming the
molecule. Most elements in the gaseous
state form diatomic molecules that consist
of two atoms of that element loosely
bound by electronic forces. The molecular
mass of diatomic gases is twice the atomic
mass. For example, the element oxygen O
has an atomic mass of 16; gaseous oxygen
O2 has a molecular mass of 32. Exceptions
to this diatomic arrangement in gases
include most metallic vapors and the
noble gases. The noble gases (argon,
helium, neon, krypton, radon and xenon)
have extremely stable electronic structures
and typically do not combine with any
other atom species. The molecular masses
of the monatomic gases are identical to
their atomic masses.
In chemical compounds containing
different elements (for example, carbon
dioxide) the molecular mass is the sum of
the atomic masses of the constituent
atoms. One carbon atom (atomic mass 12)
combines with two oxygen atoms (atomic
mass 16 each) to form CO2 with a
molecular mass of 44. The molecular
TABLE 6. Viscosity and molecular masses
of typical gases and vapors used in leak
testing.
Gas
Viscosity at
15 °C (60 °F)
(µPa·s)a
Hydrogen
Helium
Methane
Ammonia vapor
Water vapor
Neon
Nitrogen
Air
Oxygen
Hydrogen
chloride vapor
Argon
Carbon dioxide
Relative
Molecular
Mass (u)b
8.7
19.4
10.8
9.7
9.3
31.0
17.3
18.0
20.0
2.02
4.00
16.0
17.0
18.0
20.2
28.0
28.7
32.0
14.0
21.9
14.5
36.5
39.9
44.0
a. One µPa·s = 10 micropoise.
b. One unified atomic mass unit (u) ≅ 1.6605 × 10–27 kg.
masses of common gases and some vapors
are tabulated in Table 6.
Vapors resulting from evaporation of
liquid hydrocarbon compounds have
molecules containing relatively large
numbers of atoms. Molecular masses of
such organic chemical compound vapors
increase as the macromolecules increase
in complexity and contain more atoms.
Stratification of
Constituents in Mixtures of
Gases
If a tracer gas is added to air already
within a vessel or system under test, a
uniform mixture of gases is often difficult
to achieve. The tracer gas will settle
toward the top or toward the bottom of
large containers, depending on the
density of the tracer gas relative to the
density of the air or other pressurizing
gases within the system. This stratification
of mixed gases is more pronounced with
high molecular mass gases and with gases
with low diffusion coefficients.
Precautions should be taken to avoid or
correct stratification effects during leak
testing by (1) premixing of tracer gas with
diluent gases before injection, during
pressurization of the test system or
enclosing hood or chambers and
(2) providing some means for circulating
and mixing the gases within large volume
chambers or test systems.
Usually, there should be no problems
with pooling or stratification inside test
systems, if precautions are taken to mix
the tracer gas thoroughly with the diluent
gas in pressurization of the test system.
However, if the test pressure is to be about
atmospheric pressure in the test system,
the system should first be evacuated to
remove air at atmospheric pressure and to
replace it by the thoroughly mixed
combination of tracer gas with diluent
gas.
Equilibrium Distribution
Law for Gas Concentration
Ratios with Gravity Effect
The preferred technique is that in which
both the tracer and diluent gases used in
pressurization of test systems are
premixed or added simultaneously
through a screened aperture or rake so as
to be mixed rather uniformly from the
start. There should then be no problem of
pooling of denser constituents inside the
system under test, provided that
precautions are taken to mix the tracer
thoroughly with diluent gas in the
pressurization of the system. The
equilibrium distribution law of Eq. 19
Tracer Gases in Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
43
gives the ratio Ch of tracer as
concentration at the top of a tank relative
to the concentration Co of the same
gaseous constituent at the bottom of the
tank:
(19) Ch
=
−
Co e
Mgh
RT
where M is molecular mass of gaseous
constituent, h is height of interior volume
of tank, R is universal gas constant, T is
absolute temperature and g is local value
of acceleration due to earth’s gravity.
From Eq. 19, it is evident that with a
specific tracer gas in equilibrium
distribution, the concentration of tracer
gas diminishes exponentially as height
within the chamber increases. The
greatest concentration of the gaseous
constituent is at the bottom of the tank
and the lowest concentration exists at the
top of the tank. However, this effect takes
no account of the relative densities of the
tracer gas and diluent gas. If the tracer gas
were lower in density than the diluent gas
(as with helium tracer gas in air),
stratification effects could have a
predominant effect, with helium
collecting at the top of the tank after a
period of time. If the tracer gas were
higher in density than the diluent gas (as
with refrigerant-12 gas in air),
stratification effects could also
predominate and the denser tracer gas
would tend to collect at the bottom of the
tank after a period of time. In large test
chambers or enclosing hoods, it would be
desirable to provide constant internal
circulation and mixing of the internal
contents of tracer gas and diluent gas, as
with a fan.
44
Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
PART 2. Mechanisms of Gaseous Flow through
Leaks
Modes of Gas Flow
through Leaks of
Restrictions
To clarify the problem of leakage, it is
necessary to consider gas flow through
small restrictions. It is extremely
important to know something about the
basic modes of flow: viscous, transitional
and molecular. Viscous flow may be
further divided into laminar flow or
turbulent flow. Other special modes of
leakage or flow are permeation and
choked flow. The factors that influence
gaseous flow through leaks are (1) the
molecular mass of the gas, (2) the
viscosity of the gas, (3) the pressure
difference causing the flow, (4) the
absolute pressure in the system and
(5) the length and cross section of the
leak path.
An understanding of leakage
mechanisms and controlling factors is
vital to the proper interpretation of leak
tests. A simple description of gaseous flow
through leaks is presented here for leak
testing operators and supervisory
personnel, followed by a theoretical
approach to leakage.
Permeation of Gases through
Solids
Permeation is the passage of fluid into,
through and out of a solid barrier having
no holes large enough to permit more
than a small fraction of molecules to pass
through any one hole. The process also
involves diffusion through a solid and
may involve many phenomena such as
adsorption, dissociation, migration and
desorption.
The first implication of permeation is
that if the system is to be relatively
leaktight, the materials of construction
have to exclude leakage by permeability.
As an example, the permeation rate at
room temperature of a natural rubber
gasket (2.5 mm thick, with a 2.5 mm wide
rim and a 125 mm diameter) with a
100 kPa (1 atm) hydrogen pressure
differential is 1.6 × 10–6 Pa·m3·s–1
(1.6 × 10–5 std cm3·s–1). In some uses, this
permeation might represent an
unacceptable leakage rate.
Another similar example of this type of
permeation involves a rubber O-ring.
Depending of the material and the type of
gas, a rubber O-ring usually represents a
permeability of about 5 × 10–7 Pa·m3·s–1
(5 × 10–6 std cm3·s–1) for every 100 kPa
(760 torr) of pressure differential per
linear centimeter of exposed O-ring
surface. This permeability does not have
to be taken into consideration during
routine leak testing if the leakage
measurement occurs in a time too short
to permit the saturation and mass transfer
of gas through the O-ring.
Mean Free Path of
Gaseous Molecules
Molecular flow occurs when the mean
free path of a tracer gas is greater than the
cross section dimension of the leak. The
mean free path is the average distance a
molecule travels between successive
collisions with other molecules in vapor
state. The mean free path is of some
importance in leak testing because it
establishes the type of gas flow that will
occur. The mean free paths of several
gases are given in Table 4.
In flow systems encountered in leak
testing, knowing the mean free path
allows one to know, or at least estimate,
the type of flow occurring. Table 4 shows,
in general, the relationship of mean free
path to pressure and the information may
be used as a guide to determine the nature
of the flow.
Characteristics of
Molecular Flow of Gases
It should especially be noted that in
molecular flow the leakage rate is
proportional to the difference of the
pressures. Molecular flow occurs quite
often in vacuum testing applications. In
molecular flow, molecules travel
independently of each other. It is possible
for random molecules to travel from a
part of a system at low pressure to
another part of the system at a higher
pressure. When an ultrahigh vacuum
system is being tested by a mass
spectrometer leak detector, the mass
spectrometer leak detector operates at a
pressure of about 10 µPa (0.1 µtorr)
whereas the ultrahigh vacuum system
might be operating at a pressure of
Tracer Gases in Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
45
0.1 µPa (1 ntorr). When a tracer gas enters
the ultrahigh vacuum system through a
leak, it will eventually travel from the
0.1 µPa (1 ntorr) vacuum system to the
mass spectrometer operating at 10 µPa
(0.1 µtorr) by the process of molecular
flow. This does not imply that the total
flow is from a system at low pressure to
one at high pressure. The mass
spectrometer operating at 10 µPa
(0.1 µtorr) sends some gas molecules into
the system at the lower pressure.
However, when summing flows, total net
flow is from the high-pressure to the low
pressure region. The high pressure system
is contributing gas molecules to the
ultrahigh vacuum system. The tracer gas
flow in the direction opposing the major
flow of molecules is possible because of
the random mode of molecular flow. The
gas molecules, when traveling from one
system to the other, do not come in
contact with molecules traveling in the
other direction.
Characteristics of
Transitional Flow of Gases
Transitional flow occurs when the mean
free path of the gas is about equal to the
cross section dimension of the physical
leak. It occurs under conditions
intermediate between laminar and
molecular flow. The transition from
laminar flow to molecular flow is gradual.
The mathematical treatment of this
region is extremely difficult; however, a
treatment of this region is necessary
because leakage from an enclosed volume
to a vacuum necessarily involves a
transition from laminar to molecular flow.
Characteristics of Laminar
Flow of Gases
The laminar flow of a fluid in a tube is
defined as a condition in which there is a
parabolic distribution of the fluid velocity
in the cross section of the tube. The two
most important characteristics of laminar
leaks are (1) the flow is proportional to
the square of the pressure difference
across the leak and (2) the leakage is
inversely proportional to the leaking gas
viscosity. Table 1 shows that the viscosity
of most gases varies by less than one order
of magnitude. Changing the tracer gas
will not markedly increase the sensitivity
of the leak test unless this change of gas
implies a change of instrument sensitivity.
However, increasing the pressure
difference across the leak by a factor of a
little over three will increase the flow rate
through this leak by a factor of ten.
Obviously then, when the leaks to be
46
Leak Testing
measured are in the laminar flow range,
the simplest means of increasing test
sensitivity is by an increase of pressure
across the leak.
Viscous Flow of Gases
through Leaks
Laminar flow is one of the two classes of
viscous flow; the other class is turbulent
flow. Because turbulent flow is rarely
encountered in leaks, the term viscous flow
is sometimes incorrectly used to describe
laminar flow in leak testing. Viscous flow
implies that the flow occurs when the
mean free path of the gas is smaller than
the cross section dimension of the leak. It
should especially be noted that the
viscous flow leakage rate is proportional
to the difference of the squares of the
pressures. Viscous flow leakage occurs in
high pressure systems, such as are
encountered in detector probe leak tests.
It is often related with the Reynolds
number. The dimensionless Reynolds
number is the ratio of the inertial to the
viscous forces acting on the medium. In
the case of tubes (or leak paths), the
Reynolds number NRe is expressed by
Eq. 20:
(20)
N Re
=
vd
η
where v is velocity (m·s–1), d is diameter of
opening (meter) and η is kinematic
viscosity (m2·s–1). However, any set of
consistent units may be used in this
equation. Above a critical value of the
Reynolds number (about 2100 in the case
of circular tube flow), flow becomes
unstable. This results in innumerable
eddies or vortexes in the flow. The partial
path in turbulent flow leaks is very erratic.
In laminar flow, the particles flow nearly
straight line paths.
Characteristics of Choked
(or Sonic) Flow of Gases
Choked flow, or sonic flow as it is
sometimes called, occurs under certain
conditions of leak geometry and pressure.
Assume there exists a passage in the form
of an orifice or a venturi, and assume that
the pressure upstream is kept constant. If
the pressure downstream is gradually
lowered, the velocity through the throat
or orifice will increase until it reaches the
speed of sound through the fluid. The
downstream pressure at the time the
orifice velocity reaches the speed of sound
is called the critical pressure. If the
downstream pressure is lowered below
this critical pressure, no further increase
in orifice velocity can occur, with the
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
TABLE 7. Composition and partial pressures of dry air at sea level (101.325 kPa or
1.00 atm). Note the similarity of partial pressures with the percentages. When less
precision can be tolerated, use percentages × 103.
Constituent
Nitrogen
Oxygen
Argon
Carbon dioxide
Neon
Helium
Krypton
Xenon
Hydrogen
Methane
Nitrous oxide (N2O)
Content
________________________
Percent
µg·g–1
78.084
20.946
0.934
0.033
1.8 × 10–3
5 × 10–4
1 × 10–4
——
——
——
——
——
——
——
——
18.18
5.24
1.14
0.087
0.5
2.0
0.5
consequence that the maximum mass
flow rate has been reached. This condition
is known as choked or sonic flow.
Leaks Dependent on a Critical Gas
Temperature or Pressure
Both pressure dependent leaks and
temperature dependent leaks have been
observed, but in extremely limited
number. Pressure dependent or temperature
dependent leaks denote a condition where
no leakage exists until a critical pressure
or temperature is reached. At this point,
the leakage appears suddenly and may be
appreciable. Further changes in pressure
or temperature cause the leakage to vary
in the prescribed manner. When the
pressure or temperature is reversed, the
leakage follows the prescribed course to
the critical point at which leakage drops
to zero. No adequate explanation for this
phenomenon is advanced, but in view of
the very few times this occurs, such leaks
can generally be ignored. Temperature
and pressure are not normally applied in
the course of leak testing for the purpose
of locating such leaks. Instead, they are
used to force existing discontinuities to
open, so as to start or increase the leakage
rate to a point of detection.
Partial Pressure
________________________________
Pa
(torr)
7.9119 × 104
2.1224 × 104
9.460 × 104
3.34 × 101
1.84
5.3 × 10–1
1.16 × 10–1
8.8 × 10–3
5 × 10–2
2 × 10–1
5 × 10–2
(4.9343 × 102)
(1.5919 × 102)
(7.10)
(2.50 × 10–1)
(1.38 × 10–2)
(3.98 × 10–3)
(8.66 × 10–4)
(6.61 × 10–5)
(3.80 × 10–4)
(1.52 × 10–3)
(3.80 × 10–4)
be used. Table 7 lists the standard
composition of dry air at sea level.
The physical properties of gases and
vapors are also important, including the
molecular mass, the molecular diameter
and the viscosity. The gas streaming
through a narrow bore tube experiences a
resistance to flow so that the velocity of
gas flow decreases uniformly from the
center outwards until it reaches zero at
the walls. Each layer of gas parallel to the
direction of flow exerts a tangential force
on the adjacent layer, tending to decrease
the velocity of the faster moving layers
and to increase that of the slower moving
layers. The property of a gas or liquid by
virtue of which it exhibits this
phenomenon is known as internal
viscosity.
The internal viscosity is directly
proportional to the velocity gradient in
the gas. Furthermore, the viscosity must
depend on the nature of the fluid. In a
more viscous fluid the tangential force
between adjacent layers for constant
velocity gradient will be greater than in a
less viscous fluid. For any gas at constant
temperature, the gas viscosity is
independent of the pressure. However, gas
viscosity increases as gas temperature
rises. Conversely, the viscosity of all
ordinary liquids decreases with increased
temperature.
Physical Properties of
Tracer Gases Used in Leak
Testing
When performing any leak test it is
important to have some knowledge of the
residual gases present in the test area
because this will have a bearing on the
choice of tracer gas and test technique to
Tracer Gases in Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
47
PART 3. Practical Measurement of Leakage
Rates with Tracer Gases
Principles of Leakage
Measurement
Criteria to Determine Type
of Gas Flow through Leaks
All leak detection with tracer gases
involved their flow from the high pressure
side of a pressure boundary through a
presumed leak to the lower pressure side
of the pressure boundary. When tracer
gases are used in leak testing, instruments
sensitive to tracer gas presence or
concentration are used to detect outflow
from the low pressure side of the leak in
the pressure boundary.
Where leak tests involve measurements
of change in pressure or change in
volume of gas within a pressurized
enclosure, the loss of internal gas pressure
or volume indicates that leakage has
occurred through the pressure boundary.
When evacuated or low pressure test
systems or components are surrounded by
higher pressure media such as the earth’s
atmosphere, or a hood or test chamber
containing gases at higher pressures,
leakage can be detected by loss of pressure
in the external chamber or by rise in
pressure within the lower pressure system
under test.
The type of flow that occurs through leaks
depends on the factors listed earlier. In
flow systems encountered in leak testing
with gases, the length of the mean free
path of the gaseous molecules can be used
to estimate the type of flow occurring
through leakage paths. (The mean free
path lengths for various gas molecules can
be calculated by means of Eq. 15 or 16.
Tables 4 and 5 give data on mean free
path lengths for several gases and pressure
ranges.) When determining the nature of
flow of gases through leaks, use is made of
two parameters: (1) the mean free path
length λ is determined by using the
average pressure in the leak flow system.
The criteria that determine the mode of
gas flow through leaks, given in terms of
the mean free path length λ and the leak
dimensional constant d, are as follows.
Modes of Gas Flow
through Leaks
In molecular flow, the mean free path
length is greater than the largest linear
dimension of the cross section of the leak.
For each type of leak test, it is essential
that the test operator have a basic
understanding of the types of flow that
might occur in a leak. Different basic laws
relate leakage rate to pressure difference
across the leak, the range of absolute
pressure involved and the nature of the
gaseous fluid escaping through the leak.
There are three basic types of gas flow
through leaks.
1. Viscous flow typically occurs in probing
applications with gases leaking at
atmospheric or higher pressures.
2. Molecular flow usually occurs in leaks
under vacuum testing conditions.
3. Transitional flow occurs under test
conditions intermediate between
vacuum and pressures higher than
atmospheric pressure.
Figure 7 shows the range of conditions of
gas pressure and leak radius under which
each of these types of flow is typically
encountered, for leakage flow of air.
48
Leak Testing
1. When the ratio λ·d–1 is less than 0.01,
the gas flow is viscous.
2. When the ratio λ·d–1 has values
between 0.01 and 1.00, the gas flow is
transitional.
3. When the ratio λ·d–1 is greater than
1.00, the gas flow is molecular.
Relation of Viscous
Leakage Flow to Pressure
Differential across Leaks
Viscous flow occurs when the mean free
path length of the gas is significantly
smaller than the cross section of a leak.
This condition is implied by the first
criterion above, where λ is at least 100
times smaller than the leak’s cross
sectional diameter d. Viscous flow occurs
in high pressure systems such as in
probing applications where tracer gases
leak into air at atmospheric pressures.
With viscous flow through leaks, the flow
rate or leakage Q is proportional to the
difference in the squares of the pressures
acting across the leak. This relationship is
shown by Poiseuille’s law for viscous flow
through a cylindrical tube, in Eqs. 21 and
22 for the leakage rate Q:
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
(21) Q
=
πr 4
8n l
or
(22) Q
=
π r4
16 n l
(
Pa
(P
P1 − P2
2
1
2
− P2
)
)
where Q is gas flow rate (Pa·m3·s–1), r is
radius of leakage tube (meter), l is length
of leakage tube (meter), n is viscosity of
leaking gas (Pa·s), P1 is upstream gas
pressure (pascal), P2 is downstream
pressure (pascal) and Pa is average pressure
within leak path, (P1 + P2)/2 (pascal).
in the discussion above. Molecular flow
usually occurs through leaks in vacuum
systems or systems that have vacuum
applied to the lower pressure side of the
pressure boundary for purposes of leak
testing. With molecular flow through
leaks, the leakage rate Q is proportional to
the difference in pressures applied across
the leak. This relationship is shown by
Knudsen’s law for molecular flow through
a cylindrical tube, neglecting the end
effect, as shown in Eq. 23 for the leakage
rate, Q through a tubular leak with
molecular flow:
(23) Q
Relation of Molecular
Leakage Flow to Pressure
Differential across Leaks
Molecular flow occurs when the mean
free path length of the gas molecules is
greater than the largest cross sectional
dimension of a physical leak. This
condition is implied by the third criterion
=
3.342
r3
l
RT
M
(P1
− P2 )
where Q is leakage rate (Pa·m3·s–1), r is
radius of leakage tube (meter), l is length
of leakage tube (meter), M is molecular
weight of gas (kilogram per mole), P1 is
upstream pressure (pascal), P2 is
downstream pressure (pascal), T is
absolute temperature (kelvin); and gas
constant R = 8.315 J·mol–1·K–1.
FIGURE 7. Types of flow characteristics of tracer gases though leaks as function of leak channel
radius and gas pressure. Graph illustrates air at 25 °C (77 °F).
105
(4 × 103)
104
(4 × 102)
103
(4 × 101)
102
(4 ×
101
(4 × 10–1)
Radius of tube, mm (in.)
Viscous
100)
Transition
100
(4 × 10–2)
10–1
(4 × 10–3)
10–2
(4 ×
10–3
(4 × 10–5)
10–4
(4 × 10–6)
10–5
(4 × 10–7)
Molecular
10–4)
10–4
(
10–3
10–2
10–1
100
101
102
103
1.5×10–8)(1.5×10–7)(1.5×10–6)(1.5×10–5)(1.5×10–4)(1.5×10–3)(1.5×10–2)(1.5×10–1)
104
105
(1.5)
(15)
Absolute pressure, Pa (lbf·in.–2 )
Tracer Gases in Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
49
If this value is substituted for R in
Eq. 23, the leakage rate in SI units is given
by Eq. 24:
(24) Q
=
9.637
r3
l
T
M
(P1
− P2 )
If the molecular mass M is given in units
of grams per mole. All other quantities are
in SI units as listed above. The leakage
rate in SI units of Pa·m3·s–1 is given by Eq.
25:
(25) Q
=
304.8
r3
l
T
M
(P1
− P2 )
In cgs (centimeter-gram-second) units,
with tube radius and length given in
centimeter and molecular weight in gram
per mole, the rate of leakage (L·s–1) is
given by Eq. 26:
(26) Q
=
30.48
r3
l
T
M
(P1
− P2 )
This flow rate is related to the flow rate Fo
for zero thickness orifice:
(27)
F
=
8
3
r
l
where both F and Fo are flow rates (L·s–1).
Relation of Transitional
Leakage Flow to Mean
Free Path Length of Gas
and to Pressure
Differential Applied across
Leak Path
Transition flow occurs when the mean
free path length of the gas molecules is
about equal to the cross-sectional
dimension of the leak. Transitional flow
occurs under leakage conditions
intermediate between those for viscous
flow and those for molecular flow. For
transitional flow, Knudsen’s law (see
Eqs. 23 to 27) for molecular leakage is
modified by an additional term that
depends on the ratio R equal r/λ or
leakage tube radius r to the mean free
path length λ that applies for the average
pressure (P1 + P2)/2, existing within the
leakage path. This correction term for
transitional flow in leakage paths is given
as the factor FT, defined by Eq. 28 where
Rt = r/λ:
(28) F T
=
0.1472 R t
+
1 + 2.507 R t
1 + 3.095 R t
The leakage rate Q in SI units of Pa·m3·s–1
is given by Eq. 29:
50
Leak Testing
RT
M
(P1
− P2 ) F T
For Eq. 28 and 29, the symbols are
explained below Eq. 23, with the
exception of the mean free path length λ,
which is determined at the average
between upstream and downstream
pressures acting across the leak, namely
(P1 + P2)/2, from Eq. 15 or 16 and Tables 4
and 5.
Analogy between Electrical
Conductance and Gaseous
Conductance
Conductance is a term describing the
property of a gas flow system that permits
gas to flow. It is defined analogously to
electrical conductance G, the reciprocal of
electrical resistance R. Ohm’s law for
direct current flow i through a
conductance or resistance is stated in
Eq. 30:
(30)
Fo
r3
l
(29) Q = 3.342
i
=
V
R
= VG
In Eq. 30, the quantity V equals the
voltage drop across the resistance R or
conductance G. Electrical conductance
could be described as the property of an
electric circuit that permits current to
flow. In steady state direct current circuits,
the conductance G is the ratio of the
current i flowing in the resistive element
to the drop in electrical potential (or
electrical pressure) across the resistive
element, as in Eq. 31 for the electrical
conductance G:
(31) G
=
1
R
=
i
V
With gas flow through the
conductance of a leak path, for example,
the flow rate Q is analogous to the electric
current i. The pressure drop (P1 – P2) is
analogous to the voltage drop V. Leak
conductance C is analogous to the
electrical conductance G. The electrical
current could be considered as the leakage
of electrical charge through a resistive
element such as a length of wire of given
diameter and specific conductivity. The
gaseous conductance of a tubular
passageway permits the leakage of a
gaseous constituent when a pressure drop
exists between the ends of the tubular
hole. The gaseous conductance is the
reciprocal of the resistance of the leak
passageway, as indicated by Eq. 32 for the
gaseous conductance C:
(32) C
=
1
R gas
=
Q
P1 − P2
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
The equivalent of Ohm’s law for a gas
conductance would be the linear
relationship of Eq. 33 for the rate Q of
leakage or of gas flow:
(33) Q
=
P1 − P2
R gas
=
(P1
− P2 ) C
However, Eq. 33 is only true for the case
of molecular flow, as shown in Eqs. 23
to 27.
It is very important to keep in mind
that, by definition, the relationships of
gaseous conductance calculations always
include a term describing a property of
the flowing gas. This property usually is
the gas viscosity that influences viscous
flow through leaks or is the gas molecular
mass that influences molecular flow
through leaks.
The conductance for molecular flow of
gases through a long cylindrical tubular
leak channel can be calculated from
Eq. 36:
(36) C
The following equations give basic
relationships required to calculate leak
conductances under various conditions of
leak geometry and of modes of gas flow
and to estimate variations of leakage rate
with different gas pressures.
The conductance of a leak exhibiting
viscous flow of gas can be calculated by
Eq. 34, assuming that the physical leak
channel approximates a straight,
cylindrical tube: The viscous conductance
C of a tube is expressed as follows:
(34) C
=
πr4
Pa
8 nl
In Eqs. 34 through 38 for calculating
the conductance of leaks, C is gas
conductance (m3·s–1), r is radius of bore of
tube (meter), l is length of tubular leak
passageway (meter), n is viscosity of
leaking gas in Pa·s, P1 is upstream pressure
(pascal), P2 is downstream (pascal), Pa is
average gas pressure within leak channel
(pascal), Pa = (P1 + P2)/2, M is molecular
mass of gas g·mol–1 and T is absolute
temperature (kelvin).
If viscous leakage occurs through an
ideal orifice where the ratio P2/P1 of
downstream to upstream pressures is
smaller than or equal to 0.52, the
approximate conductance for viscous flow
through an ideal orifice can be calculated
by the empirical Eq. 35:
(35) C
=
6.4 r 2
P
1 − 2
P1
T
M
r3
l
3.342
RT
M
The conductance of a leak that can be
approximated by an ideal orifice subject
to molecular flow at low pressure is
calculated by Eq. 37:
(37) C
=
T
M
3.613 r 2
The conductance of a long tubular leak
with transitional gas flow can be
calculated from Eq. 38:
(38) C
Equations for Calculating
Conductances of Leaks
with Various Modes of
Flow
=
=
r3
P
3.342
T
M
FT
In Eq. 38, the factor FT is the correction
term for transitional flow defined earlier
by Eq. 28.
Gas Conductance with Two Leaks
in Series
If two different diameter leaks with
different conductance values are
connected in series as in Fig. 8, the total
conductance of the connection between
extreme ends decreases (resistance
increases). From Eq. 33, the conductance
of the leak between the outer ends of
sections 1 and 3 may be expressed as in
Eq. 39:
(39) C1 − 3
=
Q
P1 − P3
The total pressure drop across the two
leaks in series is given by Eq. 40:
FIGURE 8. Diagram of typical leak paths connected in series.
Wall of
system
P1
Inside
of
system
Atmosphere
P3
P2 chamber
C12
C23
Tracer
Inner
capillary
Outer
capillary
Legend
C = channel connecting points
P = point where fluid is present
Tracer Gases in Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
51
(40)
P1 − P3 =
(P
1
) (
)
− P2 + P2 − P3
The pressure drop across each
individual leak is shown in Eq. 41:
(41)
P1 − P2
=
P2 − P3
=
Q
C1 − 2
Q
or, in its reciprocal form,
Q
(43)
C13
=
=
1
+
1
CT
=
=
C1 − 2
+
1
C 2 −3
1
1
1
+
+ … +
C1
C2
Cn
C1 × C 2
C1 + C 2
This case applies for two successive leak
conductances connected in series. This is
analogous to the case of two electrical
resistors connected in parallel or of two
electrical conductances connected in series.
Leak Conductance for Two
Leaks Connected in
Parallel
Figure 9 shows the case of two leaks
connected in parallel. With this situation,
the total leakage through two parallel
leaks divides between the two leakage
paths from the high pressure side to the
low pressure side of the pressure
boundary. The division of flows depends
on the conductance of the individual
leaks as indicated in Eqs. 46 and 47:
52
Leak Testing
Ca ∆ P
(47) Q b
=
C b ( P1 − P 2)
=
C b ∆P
=
Ca ∆ P + C b ∆ P
∆P
Simplifying Eq. 48 gives Eq. 49:
(49) C1 − 2
=
Ca + C b
In its general form, the total
conductance for n individual leaks
connected in parallel is given by the sum
of the individual conductances as in
Eq. 50:
(50) CT
= C1 + C 2 + C 3 + … + C n
Q
where the subscript T denotes the total
conductance of a number of conductances
C1, C2, C3 … Cn connected in series.
In the case of only two conductances
connected in series, Eq. 44 should be
written in the form of Eq. 45:
(45) CT
=
Q
C 2 −3
In its general form, Eq. 43 may be written
as Eq. 44:
(44)
Ca ( P1 − P2 )
(48) C1 − 2
C 2 −3
C1 − 2
=
The total conductance through the
pressure boundary between Points 1 and 2
is given by Eq. 48:
,
Now, by combining Eqs. 39 to 41, the
conductance C13 for the two leaks in
series is given by Eq. 42:
Q
(42) C13 =
Q
Q
+
C1 − 2
C 2 −3
1
(46) Q a
Graphical Determination
of Conductance for
Molecular Flow through
Tubes and Orifices
The preceding equations give the
relationships required to calculate
conductance under various conditions. In
practice, calculation of the exact
conductance often is not required in leak
testing. Also, it has been found that most
needed conductance values are for
molecular flow through cylindrical tubing
and orifices. Figures 10 and 11 have been
provided to allow quick determinations.
Note that the curves are plotted for air at
20 °C (68 °F) and values must be corrected
if another gas or temperature is used.
FIGURE 9. Diagram of typical leak paths connected in
parallel.
Inside of
system
P1
Wall of system
Atmosphere
P2
Ca
Cb
Legend
C = channel connecting points
P = point where fluid is present
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
Improvement of Viscous
Flow Leak Test Sensitivity
by Increasing Pressure
Differential
Effect of Variations in
Tracer Gas Concentration
The change in leakage rate obtained by
increasing the pressure applied to a leak
to atmosphere is used to great advantage
in leak testing under conditions where the
leakage flow is viscous in nature, as
illustrated in Fig. 12. For example,
suppose that the leakage rate is
1 × 10–8 Pa·m3·s–1 (1 × 10–7 std cm3·s–1)
from a system with an internal gage
pressure of 1 atm (absolute internal
pressure of 200 kPa or 2 atm), as indicated
by point P1 in Fig. 12. It is desired to
determine the new absolute internal
pressure P2 needed to make the leakage
rate 50 times higher, or 5 × 10–7 Pa·m3·s–1
(5 × 10–6 std cm3·s–1). From Fig. 12, the
new flow rate at Point P2 is seen to be
obtained with an absolute internal
pressure of 1.23 MPa (12.3 atm), as shown
on the horizontal scales.
Under viscous flow conditions, which are
usually encountered when leak testing
pressurized systems, the flow rate increase
resulting from higher pressure differentials
may also be used to conserve tracer gas. In
a mixture of two gases such as helium and
nitrogen, each gas will flow through a
leak at the same rate regardless of their
concentrations in the mixture. Thus, if a
10 percent tracer gas in 90 percent carrier
gas mixture is used, the test sensitivity
will be 10 percent of what it would be if
100 percent tracer gas were used at the
same working pressure. To bring the test
sensitivity back to a leakage rate increase
ratio of 1, the pressure would have to be
raised enough to increase the flow by a
factor of 10. Suppose that a tank must be
brought to an absolute pressure of 10 MPa
(1.5 × 103 lbf·in.–2) and leak tested with
helium. To save money, it is desired (1) to
use the smallest amount of helium that
will give adequate sensitivity and (2) to
FIGURE 10. Conductance of cylindrical tubes of different lengths and inside diameters for air
at 20 °C (68 °F). 1 L = 1 dm3 = 0.028 ft3.
Conductance (L·s –1)
600
800
1000
et
s
1
er
2
200
100
80
60
40
m
(0
13
20
10
8
6
10
0.5
m
m
6
m
5
m
5
4
.)
2
.)
in
in
.)
12
.
(0
6
.1
8
(0
m
m
in
3
m
8
.1
(0
m
0.05
.2
(0
m
4
0.1
75
.3
m
0.2
5
1.0
0.8
0.6
.
in
)
0.02
0.01
0.4
0.005
0.2
0.1
0.0025
10–5
Length of tube (in.)
am s )
di be n.
e tu 4 i n.)
i
(
)
sid f
.5 n.
In o mm (3 .0 i n.)
i
0 m (3
10 m m (2.5 .)
89 m m
in
75 m
.0
.)
(2
in
62
m
.5
m
(1
.)
m
in n.)
50
m
0 5 i .)
.
(1 87 in .)
38
m (0. .75 5 in
0
m
(
2
m
25 m m (0.6 in.)
.)
22 0 m m
.5
in
2 m
(0
m
16
5
Length of tube (m)
400
20
4.0
6.0
8.0
10
40
60
80
100
1000
800
600
400
20
10
103
200
102
10
2.0
0.2
0.4
0.6
0.8
1.0
1
0.06
0.08
0.1
0.04
10–1
0.02
0.01
10–2
10–4
10–3
10–2
10–1
100
Conductance (m3·s –1)
Tracer Gases in Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
53
FIGURE 11. Conductance of orifices for air at 20 °C (68 °F), molecular flow. For curve A read
left vertical scale; for curve B read right vertical scale (1 m3 = 35 ft3).
10–1
A
B
10–3
1.0
10–4
0.1
10–5
0.01
2.5
25
(0.1)
(1 . 0 )
250
2000
(10)
(100)
Conductance (m3·s–1)
10
Rea
dr
Conductance (m3·s–1)
10–2
igh
Rea
dl
eft
s
cal
e
t sc
ale
100
Orifice diameter, mm (in.)
FIGURE 12. Viscous leakage rate as function of internal pressure of system leaking to
atmosphere when pressurizing with 100 percent gas. For curve A read left vertical scale, for
curve B read right vertical scale.
Internal absolute pressure (atm) at 1 atm
1
10
100
1000
100
100 000
P2 = 1.23 MPa (12.3 atm)
A
B
10
10 000
1000
1.0
Leakage rate increase ratio
Leakage rate increase ratio
50
P1 = 200 kPa (2 atm)
100
0.1
102
(14.7)
2×
10 2
10 3
10 4
(147)
(1470)
10 5
(14 700)
Internal pressure, kPa (lbf·in.–2), outside of part at 100 kPa (1 atm)
54
Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
pressurize the rest of the way with
nitrogen. The minimum detectable
leakage should be at least
1 × 10–8 Pa·m3·s–1 (1 × 10–7 std cm3·s–1) at
100 kPa (1.0 atm) pressure differential. It
is desired to calculate the percentage of
helium that should be used after reaching
an absolute pressure of up to 10 MPa
(1.5 × 103 lbf·in.–2).
The specified minimum detectable
leakage rate of 1 × 10–8 Pa·m3·s–1
(1 × 10–7 std cm3·s–1) requires that the leak
test sensitivity be standard or the same as
it would be if 100 percent tracer was used
at 100 kPa (1 atm) pressure difference.
From Fig. 12 it is seen that an absolute
pressure of 10 MPa (100 atm) results in a
leakage rate increase factor of 3300. Thus,
the helium concentration after pressuring
up should be 1/3300 or 0.03 percent.
Figure 13 is very similar to Fig. 12 and is
used in the same manner. The difference
is that Fig. 13 is plotted for conditions
where high vacuum is on the low pressure
side of the pressure boundary. Figure 13
still assumes viscous flow conditions.
Effect of Increasing
Pressure Differential across
Molecular Leak Flow
Figure 14 shows the effect of changing the
pressure differential across a leak when
the flow conditions are molecular. As
would be predicted by Eq. 23, the increase
of gas flow is a linear function of pressure.
Under conditions of molecular flow, the
amount of tracer gas flowing through a
leak is not a function of the total pressure.
It depends only on the partial pressure of
the tracer gas. Therefore, there would be
no advantage in raising the total pressure
difference without raising the tracer gas
pressure.
Conversions between
Leakage Rates with
Different Tracer Gases
Many occasions will arise where it will be
necessary to express a leakage rate (flow
rate) or conductance in terms of a
particular tracer gas when it has been
measured using a different tracer gas. For
FIGURE 13. Viscous leakage rate as function of pressure differential during vacuum testing,
pressurizing with 100 percent tracer gas. Read left vertical scale for curve A and right vertical
scale for curve B.
Internal absolute pressure (atm)
1
10
100
1 000
1 000 000
1 000
800
600
A
100 000
50 000
40
B
20
10
8
6
10 000
Leak rate increase ratio
100
80
60
Rea
dr
igh
t sc
ale
200
Leak rate increase ratio
500 000
Re
ad
left
sca
le
400
5 000
4
2
1 000
1
10 2
10 3
10 4
10 5
(14.7)
(147)
(1470)
(14 700)
External pressure, kPa (lb f ·in.–2 ) inside part at high vacuum
Tracer Gases in Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
55
example, a specification may state that a
certain part cannot leak more than a
given amount for air, but helium tracer
gas is used in testing. To be able to
convert a measured helium leakage rate to
an equivalent air leakage rate, the type of
flow must first be identified. After the
flow type is determined, the conversion
may be made. Also, many times the
conductance for a piece of tubing or other
item must be determined. Conductance is
the part of the flow equations that
contains the term describing either
molecular mass or viscosity of the gas that
is flowing. The conductance of a given
system will be quite different for two
gases having different properties. In leak
testing work, this situation is encountered
where the pumpdown time of a system
for air must be determined and then
response and cleanup times for helium
must be determined.
Conversion of Viscous Flow Rates
between Different Gases
If a flow rate has been identified as
viscous for one gas, the viscous flow for
any other gas may be determined using
the expression given in Eq. 51:
(51) Q 2
=
n1
n2
Q1
FIGURE 14. Molecular leakage rate as function of pressure
differential in vacuum leak testing, pressurizing with
100 percent tracer gas.
Dividing both sides of Eq. 51 by the
pressure drop will give conductance C
rather than flow Q. Any two conductances
C1 and C2 will then have a relationship
given in Eq. 52:
(52) C 2
=
n1
C1
n2
where C1 is conductance (any units) for
gas 1, C2 is conductance (same units as
gas 1) for gas 2, n1 is viscosity (any units)
for gas 1 and n2 is viscosity (same units as
gas 1) for gas 2.
A few comparisons that may be used
for converting higher conductance or flow
from helium flow rates to flow rates for
other gases are shown in Table 8.
TABLE 8. Comparison of viscous flow rates
of other gases with helium flow rates.
Q
Q
Q
Q
Q
Q
of
of
of
of
of
of
Multiply Helium
Flow by
argon
neon
hydrogen
nitrogen
air
water vapor
0.883
0.626
2.23
1.12
1.08
2.09
Conversion of Molecular Flow
Rates between Different Gases
If molecular flow occurs, the flow rate for
one gas may be compared to the flow rate
for any other gas by Eq. 53:
1000
800
600
400
200
Leak rate increase ratio
Conversion of Viscous
Conductance between Different
Gases
To Convert to
where Q1 is flow rate (any units) for gas 1,
Q2 is flow rate (same units as gas 1) for gas
2, n1 is viscosity (any units) for gas 1 and
(53) Q 2
100
80
60
40
=
M1
Q1
M2
where Q1 is flow (any units) for gas 1, Q2
is flow (same units as gas 1) for gas 2, M1
is molecular mass for gas 1 and M2 is
molecular mass for gas 2.
20
10
8
6
Conversion of Flow Rates for
Molecular Conductance
4
2
1
0.1
(1)
1.0
(10)
10
(100)
Absolute external pressure inside of part
at high vacuum, MPa (atm)
56
n2 is viscosity (same units as gas 1) for
gas 2.
Leak Testing
100
(1000)
The conductance under conditions of
molecular flow for one gas may be
compared to the conductance for another
by using the expression of Eq. 54:
(54) C 2
=
M1
C1
M2
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
where C1 is conductance (any units) for
gas 1, C2 is conductance (same units as
gas 1) for gas 2, M1 is molecular mass for
gas 1 and M2 is molecular mass for gas 2.
A few comparisons that may be used
for converting either conductance or flow
are given in Table 9.
TABLE 9. Comparison of molecular flow
rates of other gases with helium flow
rates.
To Convert to
Q
Q
Q
Q
Q
Q
of
of
of
of
of
of
Multiply Helium
Flow by
argon
neon
hydrogen
nitrogen
air
water vapor
0.316
0.447
1.410
0.374
0.374
0.469
Effect of Temperature on Gas
Conductance with Molecular Flow
The effect of temperature on conductance
when the flow is molecular should not be
overlooked. As can be seen in Eq. 35 and
36, the conductance changes in direct
proportion with the square root of gas
temperature. The expression of Eq. 55 is
for a variation in gas conductance
resulting from a change in temperature
only, with pressure and dimensions
remaining constant:
(55) C 2
=
T2
C1
T1
where C1 is conductance at temperature
T1; C2 is conductance at temperature T2;
T1 is starting temperature, kelvin; and T2
is new temperature in kelvin. T1 and T2
must be absolute temperatures.
Relative Sensitivities of
Leak Testing Techniques
When choosing a test technique it is
advantageous to have an insight into the
relative sensitivities of the various
techniques. Obviously, the test sensitivity
does not equal the published ultimate
sensitivities of the various detecting
devices because of many variables.
Table 10, showing relative sensitivities,
may be used to assist in choosing
potentially satisfactory leak testing
techniques.
Test Variables Limiting
Leak Testing Sensitivities
Some factors that prevent leak testing
devices from attaining their ultimate
sensitivities include geometry, sampling
efficiency, tracer economy and noise (or
contamination). Geometry enters the
picture because any instrument should
and must respond only to local
conditions at its sampling inlet. Two
things are of interest in the leak
evaluation process: the space coordinates
of the leaking orifice and the mass rate of
leakage. The effects of leak location and
leakage rate on the concentration of tracer
at the instrument depend on convection
and diffusion of the tracer gas.
Sampling efficiency may be thought of
both as a measure of how nearly all of the
quantity to be measured is used in
making the measurement and as a
measure of how well extraneous responses
can be excluded. Many leak detectors
must operate with their active parts in a
partial vacuum. This limits the rate at
which samples of the surrounding air can
be ingested for analysis. Other leak testing
instruments may take in the sample so
violently that extra turbulences are
created near the sampling point. The
sampling problem is somewhat
interrelated with the noise and
contamination problem.
The ultimate sensitivity of most leak
testing instruments is quoted on the basis
of 100 percent tracer concentration in the
system or, equivalently, on the amount of
tracer leaking. In a practical situation this
concentration is necessarily kept down for
reasons of safety or economy and
sometimes because of corrosiveness of the
tracer. With reduced tracer concentration,
the leakage sensitivity is reduced
proportionately. With 1 percent tracer
concentration the sensitivity figure is
correspondingly reduced by a factor of
100.
Control of Ambient
Concentrations of Tracer
Gases
Changes in tracer gas concentration due
to leaks are self obscuring in the presence
of random variations in the ambient
tracer gas concentration. Background
levels of tracer gas in the atmosphere
disturb the predicted gas concentration
pattern. The problem of distinguishing
leaks from increasing and randomly
varying background contamination may
reduce instrument sensitivities by orders
of magnitude or even destroy test
sensitivity altogether.
Tracer Gases in Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
57
TABLE 10. Relative ultimate leakage sensitivities of various leak testing methods under
ideal conditions with very high concentrations of tracer gases. (These numbers are not
intended to be used as guides in practical leak testing.)
Minimum Detectable Leakage Test Technique
Leakage Rate
__________________________________________
Pa·m3·s–1
(std cm3·s–1)
Pressure drop using liquids
Pressure drop using gases
Pressure rise
Ultrasonic leak detector
Volumetric displacement (gas flow meter)
Gas discharge
Ammonia and phenolphthalein
Ammonia and bromocresol purple
Ammonia and hydrochloric acid
Ammonia and sulfur dioxide
Halide torch
Air bubble in water
Air and soap or detergent
Thermal conductivity
Infrared
Hydrogen Pirani technique
Hot filament ionization gage
Mass spectrometer detector probe
Halogen diode detector
Hydrogen bubbles in alcohol
Palladium barrier detector
Mass spectrometer envelope
Radioactive isotopes
Depends on volume tested and gage range
Depends on volume tested
Depends on volume tested
10–2
(10–1)
10–3
(10–2)
–3
10
(10–2)
–3
–4
10 to 10
(10–2 to 10–3)
10–3 to 10–4
(10–2 to 10–3)
10–3 to 10–4
(10–2 to 10–3)
–3
–4
10 to 10
(10–2 to 10–3)
–4
10
(10–3)
10–4 to 10–5
(10–3 to 10–4)
10–4 to 10–5
(10–3 to 10–4)
10–5
(10–4)
–4
–5
6 × 10 to 6 × 10
(6 × 10–3 to 6 × 10–4)
10–7
(10–6)
10–7 to 10–8
(10–6 to 10–7)
–6
–8
10 to 10
(10–5 to 10–7)
–7
10 to 10–9
(10–6 to 10–8)
5 × 10–7
(5 × 10–6)
10–8 to 10–9
(10–7 to 10–6)
10–10
(10–9)
–9
–13
10 to 10
(10–8 to 10–12)
Any gas tracer system, no matter how
sensitive, that responds to the simple
absolute level of concentration will soon
become incapable of detecting leakage
when the ambient tracer concentration
rises to the level capable of giving
spurious signals. This is the major failing
of the simple halogen leak detector.
Two solutions to the background
problem immediately present themselves:
(1) keep the ambient concentration low
and (2) use a gradient sensor (differential
detector). One such instrument actually
has two separate detection cells
(Chambers where the temperature
compensator detects are mounted). Each
cell has an individual intake port. The
dual detectors continually compare the
thermal conductivity of the sample gas
(from potential leakage sources) with that
of the ambient atmosphere. When the
sample cell intake is not near a leak, the
two detection cells are sampling the same
gas concentration and their combined
output is zero, giving no output reading.
Only when the leak area is encountered
by the leakage sample intake does the
instrument respond.
The differential detector prevents
interference from gases in the atmosphere
and working area. It eliminates the need
for selectivity to any particular gas. Leak
testing can be performed in areas of high
58
Leak Testing
gas concentration that are caused by
accumulated leakage or by venting tracer
gases. The need for controlled
environment and ventilating systems is
minimized. The reference intake of the
differential detector is prevented from
sampling in the immediate area of the
leak to avoid fast transients and confusing
indications. However, the differential
detector leak sensor is less sensitive than
either the heated anode halogen leak
detector or the helium mass spectrometer
leak detector.
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
PART 4. Mathematical Theory of Gas Flow
through Leaks
Mechanisms of Mass
Transfer in Gas Flow
Mass transfer attributed to leakage can
occur in two modes: pneumatic flow and
permeation. Pneumatic flow occurs when
leakage is by passage of fluid through
finite holes. Permeation is passage of a
fluid into, through and out of a solid
barrier having no holes large enough to
permit more than a small fraction of the
molecules to pass through any one hole.
Leakage Rates for
Different Modes of
Pneumatic Flow of Gas in
Leaks
Pneumatic gas flow in leaks may be
placed in five categories: turbulent,
laminar, molecular, transition and choked
leakage flows. The approximate ranges of
flow rates for various pneumatic modes of
gas flow follow.
1. Turbulent flow occurs with leakage
rate above 10–3 Pa·m3·s–1
(10–2 std cm3·s–1).
2. Laminar flow occurs with leakage rates
in the range from 10–2 to 10–7
Pa·m3·s–1 (10–1 to 10–6 std cm3·s–1).
3. Molecular flow is most probable with
leakage rates below 10–6 Pa·m3·s–1
(10–5 std cm3·s–1).
4. Transition flow occurs in the gradual
transition from laminar to molecular
flow.
5. Choked flow occurs when the flow
velocity approximates the speed of
sound in the gas.
Laminar and molecular flows are the
predominant modes of leakage flow in the
range of leakage rates of interest in most
leak testing.
Because turbulent flow is rarely
encountered in leaks, the term viscous
flow is sometimes incorrectly used to
describe laminar flow in leak testing work.
The most familiar laminar flow
equation was developed by Poiseuille.
Poiseuille’s equation for laminar flow
through a straight tube of circular cross
section is given in Eqs. 21 and 22.
Poiseuille’s equation has been
substantially verified experimentally and
is applicable where the length and
diameter of the flow passage are known.
This is not the case for most leaks.
Equation 21 can be rewritten in the form
of Eq. 56:
(56) Q
=
K Pa
P1 − P2
n
K represents the constants of the two
geometry factors of length l and diameter
d of the tubular leak passage, as shown in
Eq. 57:
(57)
K
πr4
8l
=
Laminar flow takes place when the
Reynolds’ number of flow is lower than
the defined critical value. The Reynolds’
number is a unitless quantity that defines
the flow conditions and is given by
Eq. 58:
(58)
N Re
=
d ρF
n
where NRe is Reynolds’ number, p is fluid
density, n is gas viscosity, F is average flow
velocity across a plane in the tube and d is
diameter of the leak (compare with
Eq. 20).
Reynolds’ Number for Ideal Gas
By substituting the ideal gas equation (see
Eq. 9) into Eq. 58, the expression for the
Reynolds’ number for an ideal gas
becomes Eq. 59:
Characteristics of Laminar
(or Viscous) Flow
(59)
The laminar flow of a fluid in a tube is
defined as a condition where the velocity
distribution of the fluid in the cross
section of the tube is parabolic. Laminar
flow is one of the two classes of viscous
flow, the other class being turbulent flow.
where M is molecular mass, R is molar gas
constant, T is absolute temperature, Q is
leakage rate, d is leak diameter and n is
gas viscosity.
The critical value of Reynolds’ number
has been shown to depend on the
N Re
=
Q 4M
d π n RT
Tracer Gases in Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
59
LT.02 LAYOUT 11/8/04 2:14 PM Page 60
entrance conditions, roughness of the
walls of a tube and shape of the flow
path. In general, for smooth tubes with
well rounded entrances, the critical value
is about 1200.
Equation for the Viscosity
of a Gas
The kinetic theory of gases states that the
viscosity of a gas is given by the
relationship of Eq. 60:
(60)
n
=
mFa
3 2 π σ2
where Fa is a average velocity of the
individual molecules, m is molecular
mass, σ is molecular diameter and n is
viscosity of gas. The average velocity of a
gas molecule is given by Eq. 61:
(61)
F
=
8 RT
πM
The mass m of the individual molecules is
given in terms of the molecular mass M of
a specific gas:
(62) m
=
M
N
In Eq. 62, N is Avogadro’s number, i.e.,
number of molecules per mole.
Substituting Eq. 61 and 62 into Eq. 60
results in Eq. 63 for the viscosity of a gas:
(63) n
=
2 M RT
3 π3 N σ2
Equation 63 shows that the viscosity of a
gas is independent of pressure and is
proportional to the square root of
absolute temperature.
Characteristics of Laminar
Gas Leaks
The two most important characteristics of
laminar leaks shown by Eq. 21 and 22 are
(1) the flow is proportional to the
difference between the squares of the
pressures upstream and downstream of
the leak and (2) the leakage is inversely
proportional to the leaking gas viscosity.
Table 11 shows that the viscosity of most
gases is similar. Therefore, a change of
leaking gas will not markedly increase the
sensitivity of the leak testing technique
unless this change of gas implies a change
of instrument sensitivity.
However, as shown in Fig. 15,
increasing the pressure difference across
the leak by a factor of a little over three
60
Leak Testing
will increase the flow rate through this
leak by a factor of ten. Obviously then,
when the leaks to be measured are in the
laminar flow range, the simplest way to
increase leakage sensitivity is by an
increase of pressure across the leak.
Equations for the Mean
Free Path of Gaseous
Molecules
The mean free path length is the average
distance that a molecule travels between
successive collisions with the other
molecules of an ensemble. The mean free
path λ of gas molecules is given by Eq. 64:
TABLE 11. Mean free paths at 25 °C (77 °F), molecular
diameters, and viscosities for gases and vapors used in
leak testing.
Gas
Acetylene
Air
Ammonia
Argon
Benzene
Carbon dioxide
Carbon disulfide
Carbon monoxide
Dichloromethane
Ethane
Ethyl alcohol
Ethylene
Refrigerant–11
Refrigerant–12
Refrigerant–21
Refrigerant–22
Refrigerant–113
Refrigerant–114
Refrigerant–134a
Helium
Hydrogen
Hydrogen sulfide
Methane
n–Butane
n–Pentane
n–Hexane
Neon
Nitric oxide
Nitrogen
Nitrous oxide
Oxygen
Propane
Sulfur dioxide
Sulfur hexafluoride
Water
Xenon
Mean Free
Path
(mm·Pa)
Molecular
Diameter
(pm)
Viscosity
(µPa·s)
9.2
16.9
9.4
20.8
6.9
13.5
8.9
17.1
7.23
1.53
4.49
358
765
465
3.21
537
8.5
8.2
9.3
10.3
11.8
10.8
12.0
9.8
19.5
12.2
218
275
5.27
1.86
1.51
1.31
13.70
419
706
782
842
260
17.8
8.3
11.8
10.0
10.0
6.99
2.32
364
632
4.23
468
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
17.8
16.8
13.3
19.1
7.7
11.6
8.8
21.0
(64)
λ
1
=
2 π n 1 σ2
where n1 is number of molecules in 1 cm3
volume and σ is molecular diameter. The
molecular density n1 of gaseous molecules
per unit volume is given by Eq. 65:
m N
M V
=
(65) n1
where m is mass of gas, M is molecular
mass of gas, N is Avogadro’s number, i.e.,
6.023 × 1023 molecules per mole; n1 is
number of gaseous molecules per unit
volume; and V is volume containing the
gas. Replacing the volume V in Eq. 65 by
its value mRT/P from the ideal gas law of
Eq. 9 and substituting it in Eq. 64 results
in the equation for the mean free path
length λ of Eq. 66:
λ
(66)
MRT
=
2 π PN σ 2
which shows that at constant pressure the
mean free path is proportional to absolute
temperature. However, if the amount of
gas in a volume is kept constant, the
mean free path is independent of
temperature, as indicated in Eq. 64. In
Eq. 66, R is the specific gas constant and
MR equals the molar gas constant.
FIGURE 15. Relation of leakage to pressure differential with
laminar flow of helium gas in typical hardware leak.
Pressure across leak (lb f ·in.–2)
Leakage rate, Pa·m3·s –1 (std cm3·s –1)
1
10 –2
(10 –1 )
10 –3
(10 –2 )
10 –4
2
5
10
20
50
100
The molecular diameters and mean free
paths of typical leak testing gases and
vapors are listed in Table 11. As a
convenient calculation guide, the mean
free path, in meters, of air at room
temperature is given by Eq. 67:
(67)
λ air
6.8 × 10 −3
P
The pressure P is expressed in pascal in
Eq. 67 (compare with earlier Eq. 16).
Equation for Molecular
Flow of Gases
Molecular flow is flow through a duct
under conditions where the mean free
path is greater than the largest dimension
of a transverse section of the duct. In such
a flow, each atom moves independently
by random movement. Net flow is from a
volume of high concentration to one of
low concentration. The original
mathematical derivations of molecular
flow are attributed to Knudsen (see
Eqs. 23 to 27). The rate of gas flow in a
long tube is given by Eq. 68:
(68) Q
=
2 π RT
M
d3
6l
(P1
− P2 )
where d is diameter of the tube, l is length
of the tube and P2 and P1 are pressures at
the two ends. For the formula of Eq. 68 to
apply, the tube must be of a circular cross
section. For tubes and ducts of a
noncircular cross section, the conductance
is less than for tubes of circular cross
section and equal area. Equation 68
applies only if the tube is much longer
than its diameter. Any difficulty
experienced by a molecule in entering the
tube must be negligibly small compared
to the difficulty in transversing its length.
Equation for Free
Molecular Entry of Gases
into a Small Aperture
If gas molecules experience difficulties in
entering a small leak opening, the kinetic
theory shows that the rate of free
molecular escape of gas from the
container into a small aperture of area A is
given by Eq. 69:
(10 –3 )
(69) Q
10 –5
=
(10 –4 )
10
20
50
100 200
500 1000
Pressure across leak (kPa)
Legend
= Theoretical values
= Measured values
=
RT
2πM
A ( P2 − P1)
In the case of an aperture, the leak
opening does not have to be circular for
this equation to apply.
Tracer Gases in Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
61
LT.02 LAYOUT 11/8/04 2:15 PM Page 62
Flow Characteristics of
Molecular Leaks
The conductance of lines and apertures in
molecular flow is independent of pressure.
Calculations may be made of the effect of
turns, apertures and change in tube
diameter to calculate the overall flow in a
leak.
Equations 23 to 27, 68 and 69
demonstrate the general form of relations
for molecular flow through leaks. They are
not applicable in most leakage situations
because the leak length and diameter are
not known. The molecular flow of each
individual species in a gas mixture is
inversely proportional to the square root
of the individual masses. Therefore, a
certain amount of separation of gaseous
species takes place during flow through a
leak. In molecular flow, the gas molecules
travel independently of each other. Thus,
it is possible for random molecules to
travel from a part of a system at low
pressure to another part of the system at a
higher pressure.
Knudsen Equation for
Transition Flow
The transition from laminar flow to
molecular flow is gradual. The
mathematical treatment of this region is
extremely difficult, but is necessary
because a leak from a volume to a vacuum
necessarily involves a transition from
laminar to molecular flow. Equation 68
shows that the conductivity of a passage
in molecular flow is proportional to the
cube of the passage diameter and
independent of pressure. Conversely,
Eq. 21 and 22 show that the conductivity
of the same passage in laminar flow is
proportional to the pressure.
Knudsen derived a semiempirical
formula for the conductance of gas
flowing through long tubes in the
transition flow region:
(70) C
=
+
Cviscous
=
π  d
 
8  2
+
1

6
2
4
ZCmolecular
Pa
nl
RT
M
1 +
×
1 + 1.24
M
RT
d3
l
Pa
M
RT
d
n
Pa



d

n 
In this case, the gas flow rate Q =
C(P1 – P2), where C is defined by Eq. 70.
Equation 70 is valid providing that:
62
Leak Testing
1. the flow is not turbulent in any part of
the pipe and
2. the pressure difference between the
ends is not so great that the
mechanism of the flow, i.e., laminar or
molecular, changes along the pipe.
Although the first of these conditions
is usually satisfied in the leak, the second
generally is not: that is, the transition
from laminar to molecular flow does take
place within a leak. Equation 70 at low
pressures becomes an equation of
molecular flow, whereas at high pressure
this equation reduces to one of strictly
laminar flow.
Knudsen used Eq. 70 to represent his
experimental data. This equation has the
effect of molecular flow added to the
effect of laminar flow; consequently, it is
not an actual representation of the flow
mechanism taking place in the leak. The
phenomenon is better visualized by
realizing that both are occurring at the
same time.
Burrows Equation for Transitional
Flow
Burrows combined Eq. 23 to 27 for
laminar flow with that of Eq. 68 for
molecular flow to obtain the general
relation for transitional flow given in
Eq. 71:
(71) Q
=
+
π  d
 
8  2
4
2 π RT
M
Pa
nl
d3
6l
(P1
(P1
− P2 )
− P2 )
In a way, Eq. 71 accurately represents the
events occurring in the leak. Both laminar
and molecular flow always occur in a leak.
However, laminar flow is insignificant at
low pressures. The molecular flow mode
contributes little to total flow at high
pressures.
Equation 71 is not completely accurate
because of a slipping of molecules in
transition flow. In laminar flow, the
velocity of the molecular layers is
proportional to their distance from the
wall, the first layer being stationary. In
the transition region, slipping of the gas
over the walls of the tube occurs; that is,
the flow velocity at the walls is not zero.
At pressures below the viscous limit, the
slip correction becomes an appreciable
contribution to the total conductance.
With further reduction in pressure, the
dependence of flow conductance on
pressure becomes more complex. The flow
characteristics begin a progressive change
from those of viscous slip flow to those of
molecular flow, where the conductance
becomes independent of the pressure. The
complete transition from viscous to
molecular flow takes place over roughly
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
two orders of magnitude change in
pressure. This effect of slip can change the
predicted flow rate by at least 20 percent.
Because of this effect, Eq. 70 better
represents flow in the transition region
but cannot handle the total transition
region.
Other authors have attempted to derive
equations to represent this phenomenon
of transition from one type of flow to
another. One simple way is to calculate
laminar flow through one section of the
tube, calculate molecular flow through
another and approximate the region
between them.
Characteristics of
Turbulent Flow of Gases
In viscous flow above a critical value of
the Reynolds’ number (about 2100 in the
case of circular pipe flow), flow becomes
unstable, resulting in innumerable eddies
or vortexes in the flow. Any particle in
turbulent flow follows a very erratic path,
whereas in laminar flow the particle
follows a smooth line. Turbulent flow
occurs only in rather large leaks because it
requires relatively high velocity.
The laws for turbulent flow are quite
different from the laws for laminar flow.
The equation relating mass flow rate Q in
units of pressure × volume/time may be
written as Eq. 72:
(72) Q
=
π d5
(
RT P12 − P22
)
16 f Ml
The friction factor f depends on
roughness of the channel walls and can
be considered a constant in fully
developed turbulent flow.
Theory of Choked (or
Sonic) Flow of Gases
through Leaks
The phenomenon of choked flow (also
known as sonic flow) of gases is described
above. Two conditions required for
choked flow to occur are:
1. The flow passage must be in the form
of an orifice or venturi in which only
negligible fractional losses occur
upstream of the orifice or throat of the
venturi.
2. The ratio of downstream to upstream
pressure must be below a certain
critical value.
The critical ratio rc of downstream
pressure P2 of upstream pressure P1
required for choked flow is given by
Eq. 73:
(73) rc
=
P2
P1
=
 2 
 + 
1
γ
γ
γ −1
The term γ is the ratio of specific heats
defined by Eq. 75 below. The velocity of
sound through a gas can be written as in
Eq. 74:
(74)
Fc
=
2γ
γ + 1
RT1
M
where Fc is velocity of sound and T1 is
absolute temperature upstream of the
orifice where the velocity is low. The ratio
of specific heat at constant pressure to
that at constant volume is described by
gamma (γ), the ratio of specific heats
defined in Eq. 75:
(75)
γ
=
Cp
Cv
where Cp is heat capacity at constant
pressure and Cv is heat capacity at
constant volume.
The mass flow rate under a choked
flow condition is given by Eq. 76:
(76) Q
=
π d 2 P1C o
4M

2 
 RT1 γ + 
γ 1

γ +1
γ −1
where d is orifice diameter, P1 is upstream
pressure and Co is orifice discharge
coefficient.
The value of γ for an ideal monatomic
gas is 1.67. For polyatomic molecules, the
heat energy supplied is used for increasing
not only the kinetic energy of translation
but also the kinetic energy of rotation and
vibration. Because the same amount of
extra energy is required at both constant
pressure and constant volume, γ decreases
with molecular complexity. Characteristic
values of γ are listed in Table 12.
TABLE 12. Specific heats of gases at constant pressure Cp,
at constant volume Cv, and as the ratio γ of Cp·Cv–1, in
joule per mole of gas at 25 °C (77 °F) and 100 kPa
(1 atm) pressure.
Gas
Argon
Helium
Hydrogen
Oxygen
Nitrogen
Carbon dioxide
Ammonia
Ethane
Propane
Cp
Cv
Cp·Cv–1 = q
20.8
20.8
28.8
29.5
29.0
37.5
36.1
53.1
73.6
12.5
12.5
20.5
21.1
20.7
28.9
27.5
44.5
65.2
1.67
1.67
1.41
1.40
1.40
1.29
1.31
1.19
1.13
Tracer Gases in Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
63
Because of the stringent requirements,
choked flow is rarely encountered as the
predominant flow mode except in very
large leaks.
(82)
>
2100
for turbulent flow,
(83)
N Re
<
1200
Criteria for Distinction
between Modes of Gas
Flow in Leaks
for viscous flow and
Equations have been presented for the
various possible modes of flow that can be
encountered in a leak. The following rules
may be used to predict the mode most
likely to occur. In distinguishing between
laminar and molecular flow, the size of
the passage and the mean free path are
the two important parameters. The
distinction may be specified by a
dimensionless parameter called the
Knudsen number. The Knudsen number is
defined as the ratio of the mean free path
of the molecule to a characteristic
dimension of the channel through which
the gas is flowing. The Knudsen number is
defined by Eq. 77:
for either turbulent or viscous, depending
on duct conditions.
Choked flow takes place when the
pressure ratio between outlet and inlet
reaches a certain minimum value. This, of
course, depends on other characteristics,
such as aperture dimension. The formula
for the critical pressure ratio for choked
flow depends on the ratio r defined in
Eq. 73. The critical ratio below which
choked flow takes place is given by
Eq. 73. Choked flow cannot take place
when P1 is so low that molecular flow
exists.
(77)
=
NK
λ
d
where NK is Knudsen number, λ is mean
free path and d is channel diameter.
The type of flow encountered in the
various Knudsen number ranges is
described by Eqs. 78 to 80:
(78)
λ
d
<
0.01
for laminar flow,
(79)
λ
d
>
(80)
0.01
N Re
>
λ
d
>
1.00
=
d ρF
n
=
Q
d
4M
π n RT
where d is channel diameter, ρ is fluid
density, F is flow velocity, n is gas
viscosity, Q is leakage rate, M is molecular
mass, R is gas constant and T is absolute
temperature.
The distinction between laminar and
turbulent flow is shown by the numerical
criteria of Eqs. 82 to 84:
Leak Testing
<
N Re
<
2100
General Formula for
Gaseous Permeation Flow
Rate
Permeation is passage of a fluid into,
through and out of a solid barrier having
no holes large enough to permit more
than a small fraction of the molecules to
pass through any one hole. The process
always involves diffusion through a solid
and may involve other phenomena such
as adsorption, solution, dissociation,
migration and desorption. The general
formula for permeation is given by Eq. 85:
(85)
for transition flow.
Flow in the viscous region is
determined by the Reynolds’ number
described earlier in Eq. 58 and 59 and
repeated in Eq. 81:
(81)
(84) 1200
1.00
for molecular flow and
64
N Re
q
=
Kp A
∆P
l
=
(SD ) A ∆lP
where Kp = SD; q is rate of mass flow
(Pa·m3·s–1·m2); S is solubility coefficient;
D is diffusion coefficient; Kp is permeation
rate constant (per second); A is area
normal to flow (square meter); ∆P is
pressure drop along the flow path
(pascal); and l is length of flow path
(meter). The ∆P in Eq. 85 does not
represent absolute pressures, but the
difference in partial pressure of the
leaking fluid between the two sides of the
barrier.
Permeation of Helium through
Rubber
Permeation presents a problem in leak
testing equipment where the construction
materials have a high permeability to the
tracer gas. For example, if a component
containing a rubber diaphragm 1 mm
(0.04 in.) thick and 650 mm2 (1.0 in.2) in
surface area is leak tested using helium
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
gas, a leakage of about 1 × 10–6 Pa·m3·s–1
(1 × 10–5 std cm3·s–1) will be measured
across the diaphragm.
This leakage is due to permeation of
helium through the diaphragm and not to
any actual holes. It represents the
maximum sensitivity of helium leak
testing that can be performed on this
component. However, if the component is
to be used with another fluid to which
the membrane is impermeable, the
apparent leakage due to permeation
measured during the leak testing has little
meaning under operating conditions.
Another example of this type of false
reading is a rubber O-ring. Depending on
material, a rubber O-ring usually
represents a permeability of about
5 × 10–7 Pa·m3·s–1 per centimeter of O-ring
surface exposed for every 100 kPa
(14.5 lbf ·in.–2) of differential pressure.
Figure 16 is an example of the permeation
rates of O-ring of various materials. This
permeability does not have to be taken
into consideration during routine leak
testing if leakage measurement occurs in a
time too short to permit the saturation
and mass transfer of helium through the
O-ring.
Procedures for Reducing Gas
Permeability Effects during Leak
Testing
To reduce permeability as a factor in
leakage measurement, three procedures
may be used:
1. The leakage measurement may be
taken rapidly, not allowing the
FIGURE 16. Permeation rate of helium at differential pressure
of 100 kPa (1 atm) through O-rings of 4 × 4 mm (0.16 ×
0.16 in.) cross section, per 25 mm (1 in.) of length at 25 °C
(77 °F) in units of pascal cubic meter per second (left vertical
scale) and torr liter per second (right vertical scale).
Silicone (composition: 20 percent)
Permeation rate (Pa·m3·s –1)
Natural (composition: 10 percent)
10–6
Hydrocarbon (composition: 10 percent)
10 –7
Synthetic rubber (composition: 10 percent)
10–7
10 –8
10–8
10 –9
0
30
60
90
120
Time (min)
150
180
200
permeation rate (torr·L·s–1)
10–5
10 –6
material to be saturated with gas. This
is only possible if the material is
relatively thick. For example, a rubber
diaphragm will rapidly saturate and
almost immediately show leakage. On
the other hand, O-rings are relatively
thick and will not saturate rapidly
enough to give a reading within a
reasonable period of time (5 min). If
the diffusivity and solubility of the
fluid in the material are known, it is
possible to calculate the rate of
increase of leakage. However, in many
cases (where the leakage path is long),
this calculation is not necessary.
Rather than calculations, experimental
results can determine very quickly if
leakage through a thick gasket is
inconsequential for short time periods.
2. The maximum permeability of all
components and the resulting mass
transfer produced by permeability
during leak testing may be calculated
(refer to Eq. 85, below). In this way,
the permeability value will be known
and only leakage above this value will
be considered as leakage flow.
3. The last and most difficult way is to
quantitatively measure the leakage at
various pressure differentials. If gas
leaks through a hole in the
component so that the leak being
measured is pneumatic and laminar,
the flow is proportional to the square
of the pressure differential across the
leak. However, if the flow is strictly
due to permeation, then the flow
through the leak will be directly
proportional to the difference in tracer
gas concentration across the leak. In
this way, the presence of holes in the
component can be differentiated from
permeation.
General Guide to Estimating Gas
Flow Rates through Leaks
Table 10 lists the theoretical ultimate
leakage sensitivities of various leak testing
techniques under ideal conditions with
very high concentrations of tracer gas. It
is derived from the various flow equations
presented in the text. As may be seen
from Tables 8 to 10, the influence of
varying the gas is not so great as that of
varying the flow mode. Once the flow
mode is determined, the conversion to
another gas should be relatively easy to
make, providing the relationships in Table
10 are in fact correct. The major difficulty
is identifying the predominant flow
mode.
The data necessary for the conversion
of leakage rates between various gases are
relatively easy to obtain. For example, the
viscosity of many gases is published. Even
if the viscosity is not known,
approximation should not produce a large
Tracer Gases in Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
65
error. As shown in Table 11, the viscosity
of gases at constant temperature varies by
less than half an order of magnitude
between the most viscous and the least
viscous.
For molecular flow, data on the
molecular mass of the gases is easily
available and should cause no problem in
the conversion (see Table 5). If choked
flow does occur, the gamma of Eq. 75,
necessary for conversion of choked flow
leakage, is 1.67 for monatomic gases and
rapidly approaches 1 as the complexity of
the gas molecule increases.
Effect of Leak Size on Mode of
Gas Leakage Flow
By working with a variety of leaks of
different sizes and under different
conditions, some of the flow modes may
readily be eliminated. For example, if the
leakage rate is small, it is relatively easy to
assume that no turbulent flow will take
place. If the leakage goes from high
pressure to a slightly lower pressure, but
not to a vacuum, it is likely that
molecular flow is not the flow
mechanism. In this case, the flow may be
of a laminar nature and therefore
conversion to a second flow pressure is
relatively easy. Choked flow is rarely
encountered in small leaks.
Another example is that of converting
the leakage rate for gas flowing into a
vacuum to an anticipated rate for a
different pressure driving gas into the
same vacuum. If the leak is of relatively
small size, 10–6 Pa·m3·s–1 (10–5 std cm3·s–1)
or less, molecular flow will play a major
role in such a leak. However, should the
leak be relatively large, 10–4 Pa·m3·s–1
(10–3 std cm3·s–1) or greater, the leakage
will be predominately laminar. If one can
accurately predict the type of flow that
will predominate in a leak, one could
therefore make accurate conversions to a
different set of conditions. Unfortunately,
the state of the art is such that these
predictions are usually not possible.
1. If pressure is increased, correlate as
laminar.
2. If pressure is decreased, correlate as
molecular.
3. If gas is changed, correlate as
molecular.
Correlation should be performed so
that, if an error is made, actual leakage
will be no greater than that predicted in
the correlation. Correlation of leaks
resulting from increased pressure across a
leak is not recommended. An actual
measurement should be made whenever
possible to verify leakage rate.
Equation for Gas Leakage
Flow Rate in Laminar Flow
Assuming the flow mode has been
identified, the following are sample
calculations for correlation of flow rates
with the use of different gases and
pressure. The first sample calculation is
for laminar flow. The general equation for
laminar flow of gases is given by Eq. 86:
(86) Q
Many authors have predicted the following
predomination flow modes in leaks of
various sizes: turbulent flow, 10–3 Pa·m3·s–1
(10–2 std cm3·s–1); laminar flow, 10–2 to
10–7 Pa·m3·s–1 (10–1 to 10–6 std cm3·s–1);
transition flow, 10–5 to 10–7 Pa·m3·s–1 (10–4
to 10–6 std cm3·s–1); molecular flow,
10–7 Pa·m3·s–1 (10–6 std cm3·s–1) . When
there is doubt about the correctness of flow
identification, the following procedure is
recommended.
66
Leak Testing
π  d
 
8 nl  2 
4
(P1
Pa
− P2 )
where Q is leakage (mass flow in units of
pressure × [volume/time]), d is average
diameter of leak hole, P2 is pressure on
the entrance side of the leak, P1 is
pressure on the exit side of the leak,
average leak inlet and leak outlet pressures
Pa = (P1 + P2)/2, n is viscosity of the
leaking fluid or fluid mixtures and l is
leak length. Note that Eq. 86 is equivalent
to Eq. 21 given earlier.
The leak dimension of d and l are
usually not known. An apparent
conductance C may be calculated by the
formula, where this apparent conductance
is the product of π(d/2)4/8l and any unit
conversion factors. From this calculation,
an apparent leak geometry factor can be
calculated from Eq. 87:
(87) C
Estimating Mode of Gas
Leakage Flow from
Leakage Rate and Pressure
=
=
π d4
128 l
If C is calculated only for conversion
from one flow to another, the constant
does not have to be in compatible units,
providing that the same units are used
both in solving for C and using the C in
correlation equations.
Using the apparent conductance C
calculated above, the flow of any gas at
operating pressure may be predicted by
using Eq. 88:
C
P12 − P22
(88) Q =
n
(
)
A similar apparent conductance may be
calculated for other flow modes using the
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
equations given earlier in this section.
Such calculations are correct only if the
flow mode has been correctly chosen.
Categories of Anomalous
Leaks
Four types of leaks have been encountered
that do not fit in the categories already
discussed: (1) check valve leaks, (2) surface
flow leaks, (3) geometry change leaks and
(4) self-cleaning leaks.
The errors in leak measurement because of
these types of leaks could be greater than
any errors inherent in the preceding
equations for flow conversions.
Effects of Check Valve Leaks
Examples of check valve and geometry
change leaks have been found during
studies of leakage phenomena. Figure 17
is a plot of the leakage-pressure
differential obtained on a damaged needle
valve. It was observed that although the
typical laminar flow curve was obtained at
a high pressure differential, below this
pressure, the leakage abruptly stopped. On
increasing the pressure, the leak
reappeared. This phenomenon was
repeatable. This type of leak would be
particularly hard to detect because the
leak cannot be seen below a critical
pressure.
Effects of Geometry Change in
Leaks
The shape of a leak may change with
changes in system pressure. As pressure
increases, the expansion of system parts
resulting from stresses induced by the
increased pressure can cause leakage rates
of known leaks to increase beyond the
predictions of laminar flow theory. Figure
18 illustrates this increase of leakage rate
with geometry change.
Effects of Self-Cleaning Leaks
If gaskets under compression are subjected
to a high helium pressure and the leakage
rate is determined quantitatively, the
slope of the pressure leakage line is found
to be greater than two. No flow regime
would produce such a slope. However,
these curves consist of a series of lines
with a slope corresponding to that for
laminar flow.
Because the increase in leakage could
result from a permanent deformation of
the gasket, an experiment was run using
an aluminum gasket too sturdy to be
deformed. Figure 19 shows the data
obtained during this experiment. During
the original increase in pressure, the
leakage increased at a rate greater than the
FIGURE 17. Check valve leakage effect in hardware leak.
FIGURE 18. Effects on leakage of geometry changes in
gasket.
Pressure across leak (lbf ·in.–2)
1
10 –4
2
5
10
20
50
100
Pressure across leak (lbf ·in.–2)
(10 –3 )
102
103
104
Leakage rate, Pa·m3·s –1 (std cm3·s –1)
Leakage rate, Pa·m3·s–1 (std cm3·s–1)
10 –6 (10–5)
10 –5 (10 –4 )
10 –6 (10 –5 )
Leakage drops to less than
10–4 Pa·m3·s–1 (10–3 std cm3·s–1)
10 –7 (10 –6 )
10 –7 (10–6)
Broken lines indicate
theoretical laminar flow slopes
10 –8 (10–7)
10 –9 (10–8)
1
10
20
50
100
200
Pressure across leak (kPa)
500
2
5
10
20
50
100
1 000
Pressure across leak (MPa)
Tracer Gases in Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
67
square of the pressure increase. However,
on releasing the pressure, the leakage
decrease was proportional to the square of
the pressure decrease. A second increase
in pressure produced an increase that
retraced the leakage encountered during
the pressure decrease. It is believed that
the original pressure increase cleaned the
leakage passages. Further pressure cycling
did not affect the maximum leakage. This
suggests that whenever possible, leak
testing should be done at the proposed
operating pressure, in order that potential
leaks may be formed and observed.
Characteristics of
Absorbed or Surface Flow
Leaks
The flow of gases and noncondensing
vapors through fine capillaries and
micropores cannot be dealt with by
means of simple techniques analogous to
those applicable to molecular and laminar
flow. The narrow passages and large
surface areas involved cause surface
adsorption and surface flow to become
important factors.
The adsorption may be physical, where
only relatively weak van der Waals
attractions are involved. However, the
adsorption may also be regarded as
chemical. In this case, the surface of the
FIGURE 19. Leakage curves showing self-cleaning effects in
leaks.
Pressure across leak (lb f ·in.–2)
10
20
50 100 200
500 1000
Leakage rate, Pa·m3·s –1 (std cm3·s –1)
10 –5 (10 –4 )
10 –6 (10 –5 )
10 –7 (10 –6 )
10 –8 (10 –7 )
0.1 0.2 0.5
1.0
2.0 5.0
Pressure across leak (MPa)
Legend
= Pressure decrease
= Second pressure increase
= Initial pressure increase
68
Leak Testing
10
solid provides binding sites for the gas
atoms and the electronic structure of the
solid permits the formation of a
chemisorption bond. The nature of the
binding sites, the bonds between the gas
atoms and the surface all influence the
degree of surface migration of the atoms.
The flow along a fine capillary or
micropore is assumed to consist of two
mechanisms working simultaneously:
(1) molecular flow along the bore of the
capillary, whereby molecules are supposed
to collide with the wall, reevaporate and
collide with the wall again without
intermolecular collisions; and (2) surface
flow along the wall of the capillary,
whereby molecules are adsorbed and
diffuse along the surface of the wall. Both
these mechanisms promote gas flow from
regions of higher gas concentrations to
regions of lower gas concentrations.
Factors Influencing Surface Flow
of Gases
For a given set of conditions, the
proportion of molecules that follow the
mechanisms of adsorbed or surface flow
leakage depends on a variety of factors,
including (1) the sticking probability (the
probability that a molecule sticking the
surface will become adsorbed), (2) the
length of time the molecule remains
adsorbed (the mean surface lifetime of the
gas molecules) and (3) the coefficient of
surface diffusion of the gas molecules.
These features are, in turn, influenced by
other characteristics, such as the number
of sites occupied by the adsorbed
molecules or whether a complete
monolayer is involved.
The nearer the properties of a gas
approach those of a condensable vapor,
the greater the proportion of surface flow.
Therefore, a reduction of temperature or
an increase of pressure may sometimes
promote a total flow in excess of that
predicted by the laminar molecule theory.
Although the final leakage rate
achieved with a condensable gas may be
higher than predicted from flow theory,
there may be an initial delay of flow
because of condensation of the tracer gas
on the leak surfaces. This delay is
important if a tracer probe technique is
used for testing. For example, if butane, a
readily condensable gas, is used in the
tracer probe, some small leaks will be
missed because of the delay caused by the
adsorption. Two remedies can be
suggested to counter this problem: use of
a noncondensable gas and use of a
detector probe with condensable gases.
With use of a detector probe, the gas is
continually in contact with the leak and
equilibrium is established.
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
References
1. Nondestructive Testing Handbook,
second edition: Vol. 1, Leak Testing.
Columbus, OH: American Society for
Nondestructive Testing (1982).
2. Slattery, J.C. and R.B. Bird.
“Calculation of the Diffusion
Coefficient of Dilute Gases and of the
Self-Diffusion Coefficient of Dense
Gases.” AIChE Journal. Vol. 4, No. 2.
New York, NY: American Institute of
Chemical Engineers (1958): p 137-142.
Tracer Gases in Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
69
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
C
3
H A P T E R
Calibrated Reference
Leaks1
Mark D. Boeckmann, Vacuum Technology,
Incorporated, Oak Ridge, Tennessee
Charles N. Sherlock, Willis, Texas
Stuart A. Tison, National Institute of Standards and
Technology, Gaithersburg, Maryland
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
PART 1. Calibrated Reference Leaks
Terminology Applicable to
Reference, Calibrated or
Standard Leaks
Physical leaks suitable for checking leak
detector performance and leak test
sensitivity are a vital component of
instrumentation for leak testing. The
terms reference, calibrated and standard
leaks have been used in the past to
identify these physical leaks. To many
people, the term calibration implies the
existence of a universally accepted
standard such as those at the National
Institute of Standards and Technology.
The National Institute of Standards and
Technology has performed calibration of
helium leaks (capillary and permeation)
over the range of 10–14 to 10–6 mol·s–1
(2.3 × 10–11 to 2.3 × 10–3 Pa·m3·s–1) on a
routine basis. The uncertainties in leak
rate vary from less than 1 percent at
10–6 mol·s–1 (2.3 × 10–3 Pa·m3·s–1) to as
much as 5 percent at 10–14 mol·s–1
(2.3 × 10–11 Pa·m3·s–1). Additionally, the
National Institute of Standards and
Technology will calibrate leaks with other
gases over this range on a special test
basis. All of these calibrations are
performed while the gas is exhausted into
a vacuum. Leaks may also be calibrated by
commercial companies that derive their
measurement uncertainty from either of
two techniques. The first is that they
derive their measurements from leaks
calibrated at the National Institute of
Standards and Technology and perform
calibrations using a comparison
technique. The second technique uses
secondary techniques that derive the leak
rate through measurements of pressure,
volume, temperature and time with
instruments whose calibration can be
traced to the National Institute of
Standards and Technology. The
appropriate type of calibration will
depend on particular measurement
requirements including the required
accuracy, traceability or regulatory issues.
In some cases, accuracy in leakage
measurement is not of prime importance.
Rather, most practical situations require
that some particular leakage value not be
exceeded. It need only be established that
no leakage in the tested system is greater
than this allowable maximum leakage
rate. This practical approach to leakage
72
Leak Testing
specification requires some arbitrary
standard. However, if any doubt exists,
one need only reduce the leakage of this
arbitrary standard physical reference leak
by a sufficient safety factor to ensure that
test sensitivity meets the practical leakage
requirement within some estimated
confidence interval.
Classification of Common
Types of Calibrated or
Standard Physical Leaks
Calibrated physical leaks are designed to
deliver gas at a known rate. The most
common use of such leaks is in the
measurement of sensitivity of leak
detectors. However, calibrated leaks are
also used to measure the speed of vacuum
pumps and to calibrate pressure gages. A
standard physical leak makes feasible the
establishment of leakage rate
requirements for specifications. It also
provides a uniform reference standard for
calibrating leak detectors at different
locations where products are inspected.
This ensures more uniform agreement of
all tests.
Calibrated leaks may be divided into
two distinct categories: (1) reservoir leaks
that contain their own tracer gas supply
and (2) nonreservoir leaks to which tracer
gas is added during testing. Figure 1
shows a classification of physical leaks
used for reference, calibration or standard
leaks.
Accuracies of Reservoir Calibrated
Leaks
The uncertainty in the leak rate of fixed
reservoir leaks is due to a combination of
calibration uncertainty, leak rate decay
because of calibration, temperature effects
and leak instability. Of these, uncertainty
in the stability of the leak is hardest to
quantify. Changes in the leak rate may
occur in capillary leaks because of partial
blockage of the capillary. Changes in the
leak rate of glass permeation leaks may
occur because of the development of
microcracks in the glass. In general, these
leaks are more stable than leaks without
closed reservoirs, particularly for
calibrated leaks with values less than
10–9 mol·s–1 (2.3 × 10–6 Pa·m3·s–1).
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
Accuracies of Nonreservoir
Calibrated Leaks
The nonreservoir type of leak provides
only a hole or a series of holes and
passages that permit gas to pass through
at a known rate. The users of this type
calibrated leak must provide gas at a
known concentration, purity and
pressure.
The uncertainty in the leak rate of
nonreservoir leaks is due to a
combination of calibration uncertainty,
temperature effects, leak instability,
pressurizing gas purity and uncertain
measurements of gas pressures and
temperatures. Most if not all nonreservoir
leaks are physical leaks and are susceptible
to plugging. Because of this it is very
important that the input gas be free of
particulates and hydrocarbons. In
addition the output should be exposed to
as little contamination as possible,
especially when the leak is not
pressurized. Nonreservoir type leaks are
typically used for higher leak rates, greater
than 10–9 mol·s–1 (2.3 × 10–6 Pa·m3·s–1),
where the depletion rate of reservoir leaks
becomes greater than 20 percent per year.
The temperature coefficients of leaks can
be measured to account for changes in the
leak rate as a function of temperature.
Comparison of Standard Leaks
with and without Tracer Gas
Reservoirs
In proper leak testing practice, the
sensitivity of leak detectors is checked
frequently by calibrated leaks of reservoir
types with internal gas supply. For system
sensitivity checks, a calibrated leak
without a reservoir is preferable because it
closely imitates the behavior of an actual
leak in the object or system under test.
The calibrated leak without a reservoir is
open to local atmospheric pressure;
therefore, it requires no sensitivity
correction for pressure, temperature and
other environmental factors. In the tracer
probe mode of leak detection, tracer gas is
sprayed on the calibrated leak under the
same conditions that exist when the leak
detector is used to measure a leak in any
system or enclosure under test.
In the case of a leak containing a
reservoir, the measured sensitivity of the
leak detector is independent of the test
gas pressure and of the tracer gas
contamination of ambient air surrounding
the leak testing area. If the calibrated leak
is to be used for the measurement of an
absolute value, as in the case of the
calibration of a pressure gage or
measurement of the speed of a pump, a
leak carrying its own gas supply is
desirable.
Basic Categories of
Calibrated Gas Leaks
Generally, leaks may be grouped into
either of two categories: (1) leaks that
depend on the permeation of some
materials by certain gases and (2) leaks in
orifices that permit the flow of any gas
when a pressure differential is exerted
across the element.
Variation of the material composition,
the membrane dimensions and the partial
pressure differential of gas across the
element permit the attainment of an
almost infinite range of flow rates. The
temperature coefficients of the
permeation leak systems are appreciable.
This provides an additional means of
extending the flow range, particularly
when the other parameters are fixed or
limited.
Leaks that permeate through a
fluorocarbon resin membrane are also
available with properties similar to those
of gas leaks. The second category of orifice
leaks permits the attainment of a wide
range of flow rates by modification of the
FIGURE 1. Categories of artificial physical leaks commonly spoken of as “reference,”
“calibration” or “standard” leaks.
Leaks
Reservoir
Capillary
Permeation
Glass
Fluorocarbon
Fixed
resin
value
Fixed
value
Nonreservoir
Variable
value
Porous
plug
Porous
plug
Capillary
Fixed
value
Variable
value
Variable
value
Calibrated Reference Leaks
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
73
element dimensions and the pressure
differential across the element.
Temperature is not as great a factor
because the temperature coefficients are
small with glass orifice standard leaks.
Properties Designed in Calibrated
Gas Leaks
The range of possible flow rates of
calibrated leaks is rather severely limited
by practical considerations in the
selection of parameters for the
construction of leaks for quantitative
standards. An ideal calibrated leak should
have the following properties.
1. The leakage rate should be constant
and should remain unaffected by
ambient conditions.
2. The calibration should be accurate.
3. The physical size should be
convenient.
4. The calibrated leak should not be too
delicate or fragile.
5. The calibrated leak should have its
own gas supply.
Temperature Coefficients of
Calibrated Leaks
Unfortunately, those parameters useful in
extending the possible range of flows are
not conducive to constancy. The high
temperature coefficients of the membrane
leaks are particularly disturbing when the
changes in ambient temperatures are
frequent and there is no way of
determining whether or not the
equilibrium flow rate is reached at any
one temperature. Even the relatively small
temperature coefficients of orifice leaks
are appreciable when the temperature
varies over wide ranges.
The National Institute of Standards and
Technology measures the temperature
coefficients of leaks as a normal part of
their calibration service over the range of
0 to 50 °C (32 to 122 °F). Some
manufacturers of calibrated leaks may also
be able to measure temperature
coefficients. Normally manufacturers
assume a linear temperature coefficient of
3 to 4 percent per 1 °C (2 °F) for glass
helium permeation leaks. For the lowest
uncertainties the temperature coefficients
should be measured.
Size, Weight and Portability of
Calibrated Gas Leaks
The convenience of the physical size is a
property that would vary considerably,
depending on the use to which the
calibrated leaks is applied. In general,
complete and convenient portability of
standard leaks is desirable and is usually
available in nonreservoir standard leaks.
Portability is easily attainable with
74
Leak Testing
reservoir standard leaks with low leakage
rates of the order of 2 × 10–7 Pa·m3·s–1
(2 × 10–6 std cm3·s–1) or less.
During manufacture of calibrated leaks,
additional effort and weight can extend
the upper limit of flow by as much as a
factor of 50 without allowing the
depletion of the gas supply to cause a
falloff in leakage rate greater than
10 percent per year. Greater increases of
the upper limit call for nearly linear
increases in volume of the leak gas
reservoir and even greater increases in
weight. These reduce the portability and
ease of installation of standard leaks.
Limitations of Flow Rate
Calibration of Standard Gas Leaks
The lower limit of flow rate that is
practical for direction calibration is about
10–11 Pa·m3·s–1 (10–10 std cm3·s–1). The
degassing of the system becomes a
problem as the size of the leak is decreased.
In this range the changes of both true
leakage and virtual leakage caused by
pressure increase are nearly equal.
Two indirect techniques may be used,
either separately or in combination, to
calibrate with reasonable accuracy in the
low ranges; both techniques have been
experimentally justified. The calibration
may be made by comparison with a
standard of greater flow rate by means of
a mass spectrometer. The actual rate is
extrapolated (assuming linear response of
the instrument). Alternatively, the leakage
rate may be increased in a manner in
which the response is predictable (i.e., the
pressure response of membrane leaks is
linear) and calibration made at the higher
flow rate.
Limitations of Glass Membrane
Standard Gas Leaks
Construction of all-glass membrane leaks
that vary in flow range at ambient
temperatures from 0 to 50 °C (32 to
122 °F) is rather simple. Larger flows
require either higher pressures or
modification of membrane parameters
that tend to make them excessively
fragile.
It is possible to combine a number of
the large leak elements in parallel to
obtain greater flow when necessary.
Advantage has been taken of the relatively
sturdy nature of glass tubing of very small
cross section and correspondingly thin
walls. Elements have been made using
literally miles of such tubing in systems
designed for use at relatively high
temperatures to separate low
concentrations of helium from natural
gases. However, these elements do not
seem suitable for use under high vacuum
conditions.
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
Limitations of All-Glass Orifice
Standard Leaks
All-glass orifice leaks are more difficult to
produce with flow rates smaller than
5 × 10–9 Pa·m3·s–1 (5 × 10–8 std cm3·s–1)
unless precautions are taken to maintain
the pressure differential significantly
positive in the downstream direction
while the upstream pressure is made
subatmospheric.
The upper limit in flow rate is
determined mainly by the maximum
acceptable physical size. Glass reservoirs
become bulky when they are of adequate
size to supply a leak of the order of
1 × 10–6 Pa·m3·s–1 (1 × 10–5 std cm3·s–1)
without having the leakage rate fall off
more than 10 percent per year.
Improving Calibrated
Leakage Rate Stability by
Increasing Envelope
Pressure
Stability of leakage rate may be improved
greatly without sacrificing compactness by
enclosing the leak element in a metal
envelope and filling the envelope to a
significantly greater pressure. Membranes
that leak 5 × 10–8 Pa·m3·s–1
(5 × 10–7 std cm3·s–1) at a pressure
differential of 100 kPa (1 atm) will raise
their leakage 20× to a rate of
1 × 10–6 Pa·m3·s–1 (1 × 10–5 std cm3·s–1)
when used with a partial pressure
differential of 2 MPa (20 atm).
The leakage rate will fall off one
twentieth as much as that of a membrane
that will leak 1 × 10–6 Pa·m3·s–1
(1 × 10–5 std cm3·s–1) at atmospheric
differential, with the same volume
reservoir. Maximum envelope membrane
leak pressures are limited by their nature
to not more than 2.8 MPa (400 lbf·in.–2
gage). Orifice leaks have been used with
maximum pressures in the envelope of
12 MPa (1700 lbf·in.–2).1
Basic Characteristics of
Membrane Standard Leaks
Membrane standard leaks share several
characteristics.
1. They are restricted to usable gases,
even at elevated temperatures.
2. They have relatively high temperature
coefficients.
3. They are relatively fragile when
constructed with glass.
4. They normally have a response linear
with respect to reservoir
concentration.
5. They are almost impossible to plug.
Basic Characteristics of
Orifice Standard Leaks
The orifice standard leaks share several
characteristics.
1. They may be used with almost any gas
under conditions sufficiently removed
from liquidus conditions.
2. They have relatively low temperature
coefficients.
3. They are relatively sturdy, being able
to stand high pressure differentials, in
excess of 10 MPa (100 atm).
4. They have pressure responses that vary
from linear response for very small
leaks, about 1 × 10–9 Pa·m3·s–1
(1 × 10–8 std cm3·s–1), to direct
proportion to the square of the
pressure for very large leaks, about
5 × 10–4 Pa·m3·s–1
(5 × 10–3 std cm3·s–1).
5. They are subject to plugging by solids
or by condensation of vapors of
materials close to liquidus conditions.
Precautions with
Calibrated Gas Leaks
For maximum accuracy in the use of
calibrated leaks, the following precautions
should be taken.
1. Leakage rates should be defined as
mass units per unit time. When
volume units are used, they must be
defined by specification of the
temperature and pressure conditions
under which they are to be measured.
2. The temperature at which the
calibration is made and the
temperature at which the calibrated
leak is used should be specified. If they
are not identical, the temperature
coefficient should be used to correct
the leakage rate. For best results the
leak should be calibrated and used
under constant temperature
conditions.
3. A considerably higher than ambient
temperature surrounding the element
of the orifice leaks will tend to
decrease the possibility of plugging by
condensation of liquid.
4. If a leak is not equipped with an
integral gas supply, care should be
taken to use dry gas with orifices and
to maintain a positive pressure
differential across the element in the
downstream direction if possible.
Membrane leaks should be given
adequate time to reach an equilibrium
rate if the partial pressure differential
of the tracer gas is changed.
Calibrated Reference Leaks
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
75
leakage temperature coefficient is large
(three percent or more per degree kelvin).
Design and Construction
of Permeation Physical
Leaks
Permeation Leak for Helium
Tracer Gas
Permeation leaks use the principle of gas
diffusion through a thin wall. Tracer gas
permeates from the high leak reservoir
concentration through the wall to air or
vacuum. Leakage is governed by the
permeability of the thin membrane. The
major advantage of permeation leaks is
that they deliver extremely small
quantities of gas. The commercially
available helium leak standard range
extends from 10–7 to 10–11 Pa·m3·s–1 (10–6
to 10–10 std cm3·s–1). Because a long period
of time is necessary to achieve permeation
equilibrium, these leaks usually come
with a self-contained gas supply. However,
at small leakage rates, the leakage remains
constant over a long period of time. The
two disadvantages of calibrated
permeation leaks are (1) that they can
only be made for gases that permeate
through membranes and (2) that their
A common helium permeation leak is
shown in Fig. 2. The helium permeation
leak consists of a small helium filled metal
or glass cylinder with an integral glass
membrane at one end. Helium diffuses
through this glass at a measurable rate.
Each leak should be calibrated and labeled
with the following information: (1) name
of manufacturer, (2) model number,
(3) type of leak (glass permeation, orifice
etc.), (4) serial number, (5) composition of
fill gas, (6) leak rate, (7) calibration
temperature, (8) estimated uncertainty of
leak rate, (9) date of calibration,
(10) temperature coefficient and
(11) reservoir pressure, date of fill and
estimated depletion rate.2
The leak may contain two valves: a
vacuum valve downstream of the leak
element and a pressure (or reservoir
valve). The reservoir valve is used for
FIGURE 2. Helium permeation leak with self-contained reservoir: (a) photograph of standard helium leak and
cut away model; (b) schematic cross section.
(a)
(b)
63 mm
(2.5 in.)
maximum
Standard
vacuum
coupling
Helium
reservoir
Permeable
glass/quartz
membrane
38 mm
(1.5 in.)
outside diameter
Filling
port
32 mm
(1.26 in.)
Leak shutoff
valve
280 mm
(11.0 in.)
76
Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
refilling of the leak reservoir. The vacuum
valve is used for briefly shutting off the
helium flow for purposes of zeroing a
helium leak detector during the process of
calibration. The vacuum valve should not
be shut off for extended periods of time
(greater than 10 min) or the stability of
the leak may be affected severely.
Porous Plug Calibrated Leaks
Providing Molecular Flow of Gas
Porous plug calibrated leaks are not
commercially available but have
frequently been cited in literature. They
consist of a metal, ceramic or glass plug
containing extremely fine pores. The
major advantage of this type of calibrated
leak is that molecular gas flow occurs
through the plug. Therefore, the change
of leakage flow resulting from a change of
tracer gas can be calculated from the
kinetic theory of gas flow. Porous plug
leaks can be either reservoir or
nonreservoir type, with the choice of
materials cited above.
changed to change the leakage rate at
which tracer gas flows out of the physical
reference leak.
Variable Value Orifice Physical
Reference Halogen Leaks
The variable value physical halogen vapor
leak shown in Fig. 4 is available for
different ranges, such as 10–5, 10–6, 10–7
and 10–8 Pa·m3·s–1 (10–4, 10–5, 10–6 and
10–7 std cm3·s–1). A schematic flow
FIGURE 3. Reservoir variable rate physical orifice leak standard
(top) and fluorocarbon resin permeation leak standard
(bottom) for calibration of detector probe instruments.
Design and Characteristics
of Capillary Calibrated or
Standard Physical Leaks
Another type of commercial calibrated
leak is a single orifice in heat resistant
glass or metal, encased in a stainless steel
fixture. Tracer gas leaks through the
orifice at the rated leakage, when the leak
is placed under a specified gage pressure
(relative to atmospheric pressure). Such
capillary leaks are available in two types,
fixed value leaks and variable value leaks,
as next described.
Fixed Leakage Value Orifice
Capillary Leaks
Capillary type calibrated leaks are made
from constructed glass tubing or collapsed
thin metal tubing. These orifice leaks can
be produced from large sizes down to
about 10–8 Pa·m3·s–1 (10–7 std cm3·s–1).
Although smaller leaks of this nature can
be made, they become extremely difficult
to handle because of leak clogging.
Capillary leaks can be calibrated to deliver
one or a variety of tracer gases. Some leaks
are to be used with an independent tracer
gas supply, i.e., they simply consist of a
capillary leak attached to the system
under test. In the tracer probe method of
leak testing, tracer gas is simply sprayed
over the capillary. Alternatively, a physical
reference capillary leak can be made with
a self-contained gas supply that can be
permanently attached to the leak. Figure 3
shows a physical capillary orifice leak
with its own tracer gas reservoir and a
leak factor gage. The gage pressure may be
FIGURE 4. Variable leak rate halogen refrigerant leak standard
with physical (capillary) leak element: (a) photograph of leak
standard with internal reservoir of refrigerant gas and
another of refrigerant liquid and (b) schematic flow
diagram.
(a)
(b)
Gage
(Pressure
increase)
Liquid
reservoir
Calibrated
leak
Vapor
reservoir
Vapor
reservoir
fill valve
Detector
probe
Vent (pressure decrease)
Fill valve for
liquid halogen/refrigerant
Calibrated Reference Leaks
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
77
diagram of its system is shown in Fig. 4b.
This leak contains a reservoir of liquid
Refrigerant-134a halogenated
hydrocarbon tracer to be valved into a
ballast tank in gaseous form. Also
connected to the ballast tank is a glass
capillary tube and pressure gage. The rate
of gas leakage through the calibrated leak
depends on the pressure in the ballast
tank. Laminar gas flow occurs through the
leak. This permits the pressure gage to be
marked in leakage units, where leakage is
proportional to the ratio of the difference
between the squares of the absolute
pressures. The halogen leak standard is
commonly used with heated anode
halogen leak detectors. It is an excellent
leak standard to use with probe
instruments because the probe may be
passed directly across the leak exit. The
calibration then approximates detector
probe operating conditions.
Variable Leak Rate Helium
Reference Leaks
Variable reference leaks have been
designed to leak helium for use with
helium mass spectrometer leak detectors
equipped with a detector probe. The leak
arrangement is shown in Figs. 3 and 5.
The leak standard in Fig. 5 uses either a
capillary or fluorocarbon resin permeation
membrane as the gas flow restriction (the
time response for a glass membrane at
FIGURE 5. Variable rate helium leak standard
(capillary style sniffer).
Fill and flush
valve
6 mm (0.25 in.)
male pipe thread
Gas reservoir
Capillary pinpont
helium source
6 mm
(0.25 in.)
78
Leak Testing
room temperature would be much too
slow). The leak is designed to yield a
point source of helium to simulate a pin
hole leak. The point source of leakage
may be used to calibrate the detector
probe of a helium mass spectrometer leak
detector. The helium detector probe
calibrator may also be used as a training
tool to train operators on the distance and
speed a probe must be from a certain size
leak to detect the leak. The major
disadvantage of the helium detector probe
calibrator is that it requires an external
tank of helium for refilling, unlike
refrigerant calibrators that can store extra
refrigerant in an on-board liquid tank
(Fig. 4).
Other types of variable value reference
leaks are controlled by elegant needle
valve or crushed tubing whose
conductance is changed by flexing.
Although the conductance of these leaks
can be made quite repeatable, they should
not be considered calibrated leaks because
of a complete lack of standardization of
leakage rates in these artificial orifice
types of physical leaks.
Sources of Inaccuracy of
Leakage Measurements
with Standard Leaks
The inaccuracy of leak detector
measurements made with physical
standard leaks can be caused by factors
such as: (1) inaccuracy in calibrating the
leak, (2) nonlinearity of the leak detection
instrument, (3) variation in pressure
differential applied across the leak,
(4) impurity of gas applied to the leak and
(5) variation in the amount of gas
reaching the detector.
Accuracies of Calibrations of
Commercially Available Physical
Reference Leaks
Beginning in 1987, the National Institute
of Standards and Technology established a
leak calibration program that calibrated
leaks over the range of 10–11 to
10–3 Pa·m3·s–1 (10–10 to 10–2 std cm3·s–1).
In a 1980 study, tests of standard leaks
from various manufacturers have shown
that their accuracies could differ by more
than ±50 percent of a mean value.1 This is
shown by the experimental plot of Fig. 6,
which shows the calibrated leakage
reading compared to the response of a
linear mass spectrometer. The straight line
drawn in the graph is the least mean
square value of leakage as a function of
spectrometer response. This line does not
imply the correct value, but the general
pattern around which the values of the
leaks congregate.
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
Even the leaks made by any single
manufacturer vary by about 10 percent.
This is usually the guarantee that is
presented on purchase of the permeation
calibrated leak. Leaks of a variable type,
such as that shown in Fig. 4, are claimed
to be accurate only to ±20 percent.
Beginning in 1987, many
manufacturers of leaks began deriving
their measurements directly from leaks
calibrated by the National Institute of
Standards and Technology. The existence
of national standards in leak
measurement should improve the relative
agreement between the manufacturers of
leaks and may also reduce the
uncertainties that manufacturers provide
for calibrated leaks.
Errors in Response of Commercial
Electronic Leak Detectors
Most commercial leak detectors display
the response to a detected leak as a
current reading on a sensitive
microammeter. It is usually assumed by
the operator that current reading twice
the magnitude of a previously observable
one represents a leak of twice the size.
This assumption of linearity in response is
not necessarily correct; nonlinearity may
result from the structure of the pumping
system, the background usually associated
with the leak testing practice, the
electronic circuitry associated with the
detection system and the mode of gas
flow through the leak.
10–2
(10–1)
10–3
(10–2)
10–4
(10–3)
10–5
(10–4)
10–6
(10–5)
10–7
(10–6)
10–8
(10–7)
10–9
(10–8)
10–10
Effect of Barometric
Pressure on Leakage
Measurements
Leakage depends on the pressure
differential acting across the leak. When
leak detection is done by a tracer probe,
the pressure differential is usually 100 kPa
(1 atm). The gas is sprayed over the
suspected area without aid of additional
pressure. Should leak detection be
performed at high altitudes, the
atmospheric pressure is less than 100 kPa
(1 std atm). The magnitude of this
reduction is as much as 20 percent in
places such as Boulder, Colorado. If the
leaks that are being located are of a
laminar nature, the laminar flow through
FIGURE 7. Typical range of error possible in actual leakage
measurements with a leak detector. Variations with
magnitude of leakage increase the difficulties of correlating
measured leakage rates with standard reference leaks.
Actual leakage (relative units)
Stated leakage, Pa·m3·s–1 (std cm3·s–1)
FIGURE 6. Comparison of leakage values for leaks supplied by
various vendors, measured by linear mass spectrometer.
Resulting ion current depends on mass spectrometer
configuration.
A typical leak detector response error
curve is shown in Fig. 7. The instrument
response is not linear with leakage. This
error is added to the error that occurs
because of the difference in leak
calibration. Because of this lack of
linearity, the farther apart the two leaks
are in nominal value, the greater the error
in the calibration. Because such deviations
exist it is recommended that, when the
leakage measurement is done to a
specified high tolerance, a calibrated leak
to the exact specified value be used as a
standard.
Possible variation
in measurement
Measured
leak
(10–9)
10–14
10–13
10–12
10–11
10–10
Ion current (A)
10–9
10–8
10–7
Measured leakage
(relative units)
Calibrated
leak
Legend
= range of deviation due to nonlinearity of instrument response
= error due to comparison and instrument linearity
Calibrated Reference Leaks
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
79
the leak is proportional to the square of
the pressure differential. The values
obtained for leakage readings at the
altitude of Boulder, Colorado, are
40 percent less than those obtained with a
100 kPa (1 std atm) pressure during use.
Therefore, a leakage rate measured to
atmosphere in Boulder, Colorado will be
only 60 percent as large as with the same
leak measured at Cape Kennedy on the
seashore of Florida.
Certain calibrated leaks contain their
own gas supply, whereas others have the
tracer gas sprayed onto the entry orifice of
the leak at 100 kPa (1 atm) pressure
during use. Calibrated leaks with a self
contained gas supply always deliver to the
detector a fixed amount of gas that can be
used to measure the sensitivity of the leak
detector. On the other hand, leaks where
gas is added during use produce the
calibrated amount of leakage only when a
100 kPa (1 atm) pressure differential is
supplied. These nonreservoir physical
reference leaks therefore deliver less than
the calibrated amount of leakage when
used at high altitudes where the
atmospheric pressure is lower. However, at
these altitudes, the pressure of the tracer
gas across the leak is lower. In such cases,
a physical reference leak without its own
gas supply describes more accurately the
sensitivity of the leakage test. It is this test
sensitivity that is important in practical
leak testing.
Effect of Tracer Gas Purity
on Accuracy of Leakage
Measurements
Another source of inaccuracy is the
impurity of the tracer gas used for leakage
measurement. If a tracer probe technique
of leak location is used, the gas is sprayed
over the suspected area in the
environmental atmosphere. In such a
case, it is quite possible that the tracer gas
is diluted with air as it approaches the
leak. Therefore, the response of the leak
detector operating on the internal
vacuum of the test system will be reduced
by the amount air impurity entering the
detector with the tracer gas. In this case, a
calibrated leak with a self-contained gas
supply is undesirable because it would not
reproduce the leakage measurement
technique. In other words, the gas should
be sprayed onto the calibrated leak in the
same manner as onto the tested leak. The
gas in a self-contained calibrated leak
would be purer than the gas encountered
by simple spraying from a tracer probe.
80
Leak Testing
Effect of Position of Calibrated
Leak on Test System
Tracer gas may be absorbed on test system
surfaces as it travels to the detector. This
would decrease the response of the leak
detector. Therefore, calibrated leaks
should be positioned on the system as
near as possible to suspected leak sites to
improve accuracy. Alternatively, they may
be positioned as far away from the
detector as possible to show minimum
sensitivity. Both of these positions are
conservative choices that ensure that
leakage from test object discontinuities
will not be underrated.
Specifying Maximum Allowable
Leakage Rate
Because of the variations discussed here,
the accuracy of any leakage measurement
probably varies from half to twice the
actual value. This implies that, if a leak is
measured as 1 × 10–6 Pa·m3·s–1
(1 × 10–5 std cm3·s–1), the actual value of
this leak is between 2 × 10–6 and
0.5 × 10–6 Pa·m3·s–1 (2 × 10–5 and
0.5 × 10–5 std cm3·s–1). Therefore, if the
maximum allowable leakage rate of a
particular system is 2 × 10–6 Pa·m3·s–1
(2 × 10–5 std cm3·s–1), the specification
may be written with a leakage tolerance of
1 × 10–6 Pa·m3·s–1 (1 × 10–5 std cm3·s–1),
knowing that the accuracy of the leakage
measurement is a factor of two. There is
reasonable assurance that if the measured
leakage is not higher than that stated on
the specification, 1 × 10–6 Pa·m3·s–1
(1 × 10–5 std cm3·s–1), the actual system
leakage will be no greater than the
allowable rate, 2 × 10–6 Pa·m3·s–1
(2 × 10–5 std cm3·s–1). This technique of
specifying leakage is much more sensible
than specifying a slightly higher leakage
value, such as 2 × 10–6 Pa·m3·s–1
(2 × 10–5 std cm3·s–1), and thereby
requiring an unreasonably high accuracy
(such as ±10 percent) during leak testing.
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
PART 2. Operation of Standard (Calibrated)
Halogen Leaks
Functions of Known
Leakage Standards
Halogen Leak Calibrator
without Reservoir
The halogen gas leak detector (known also
as the alkali ion diode halogen leak
detector) is a transfer agent or compactor.
Leak testing with halogen tracer gas
requires use of a known reference halogen
leak to calibrate the leak testing operation
properly. The halogen leak detector is
adjusted to produce an alarm or meter
indication of the panel indicator when
exposed to a known leakage rate. The
detector is then used to compare
unknown leakage rates to the specific
known leakage rate of a calibrated
reference leak.
The maximum acceptable leakage rate,
however, must first be determined, either
by the user or from specifications that the
user must meet. The type and range of
leak standard then may be selected, but
only after this has been accomplished.
Three types of halogen leak standards are
(1) the calibrated standard leak (no gas),
(2) the leak capsule (single gas reservoir)
and (3) the halogen leak standard (reserve
gas supply).
Figure 8 shows a leak calibrator that has
no reservoir for halogen tracer gas. It
contains a single orifice in heat resistant
glass. When a reservoir of refrigerant-134a
is attached, the pressure of the
refrigerant-134a gas is 165 kPa gage
(24 lbf·in.–2 gage) and that gas will leak
through its orifice at a fixed rate.
FIGURE 8. Calibrator for halogen leak standards with small
bore capillary tube orifices for leaks from 3 × 10–5 to
3 × 10–8 Pa·m3·s–1 (3 × 10–4 to 3 × 10–7 std cm3·s–1) or
with larger bore capillary tube orifices for leaks from
3 × 10–4 to 3 × 10–7 Pa·m3·s–1 (3 × 10–3 to
3 × 10–6 std cm3·s–1). Movement of a colored liquid within
the calibrated capillary tube over a specific period of time
permits calculation of the rate of leakage from the standard
leak, when the calibrator is attached to the standard
through a vent valve.
Calibrated Halogen Leak
with Gas Reservoir
The calibrated halogen leak of Fig. 3 has
its own refrigerant-134a reservoir plus a
leak factor gage. The gage reads in
multiplying factors, used when the
pressure is changed to vary the leakage
rate. The gage is set at a factor of 1 at the
factory (165 kPa or 24 lbf·in.–2 gage).
These leak capsules, used when a precise
leakage rate is required, are frequently
mounted in the halogen leak detector
control unit.
Adjustable Halogen Leak
Standards with Ballast
Tank
The halogen leak standard shown in
Fig. 4a contains a reservoir of liquid
refrigerant-134a, which is valved in
gaseous form into a ballast tank.
Connected to the ballast tank is a glass
capillary tube and pressure gage. The
amount of leakage is dependent on the
amount of refrigerant-134a tracer gas
pressure in the ballast tank. Pressure is
indicated by a Bourdon gage and
controlled by two valves (Fig. 4b).
Applications of Calibrated
Halogen Leaks and
Capsules
Predetermined standard halogen leaks are
of great advantage to quality control
engineers in refrigeration, air
conditioning and space vehicle
Calibrated Reference Leaks
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
81
manufacturing, where critical checks of
lines, valves and hydraulic systems are of
the utmost importance. They afford great
accuracy wherever halogen leak detectors
are used and where a leak of one specific
value is required.
The adjustable halogen leak standard
in Fig. 4 provides the same advantages as
the calibrated leaks and capsule but has
the additional advantage of being
adjustable to the full scale rating. Thus,
they can be used more easily for
quantitative measurements of actual leaks
and of background contamination. Leak
standards also enable the establishment of
leakage rate specifications and provide
uniform standards for calibrating leak
detectors at each location of product
inspection.
Halogen Leak Standards to
Prolong Life of Alkali Ion Diode
Sensing Element
All three leak standards can be used to
extend the useful life of the alkali ion
diode sensing elements in heated anode
halogen leak detectors. Users frequently
replace the detector’s sensitive element
long before the end of its useful life. A
sensing element can be used until it no
longer responds to the desired setting of
the leak standard. Additionally, any leak
standard permits use of the lowest
possible anode heater current to provide
adequate leak detector sensitivity. This
practice increases element life and results
in reduced maintenance and lower
replacement costs.
Accuracy of Adjustable
Halogen Leak Standards
Typical accuracy of the adjustable halogen
leak standard of Fig. 4 is about ±20
percent of scale setting on the upper two
thirds of the scale and ±30 percent of
scale setting on the lower one third of the
scale.
Description of Adjustable
Halogen Leak Standard
The leak standard of Fig. 4 is a simple,
accurate instrument that expels a halogen
compound gas, refrigerant-134a through a
glass capillary marked probe to the
atmosphere at a known rate. This known
rate is adjustable when using certain
halogen leak standards. The leak standard
is intended primarily for use with halogen
sensitive leak detectors. The leakage rate
for each unit is marked on the scale plate.
Leakage rates are customarily labeled in
units of standard cubic centimeter per
82
Leak Testing
second (std cm3·s–1) and also in ounce per
year (oz·yr–1) by the manufacturer of these
standard leaks. The SI units are mole per
second (mol·s–1) and pascal cubic meter
per second (Pa·m3·s–1).
Components of Adjustable
Halogen Leak Standard
The adjustable halogen leak standard is a
compact instrument consisting of the
following seven functional components
(see Fig. 4):
1. direct reading leakage rate indicator
(calibrated in ounces of
refrigerant-134a per year);
2. probe fitting in the center of which is
a glass leak capillary (a different
capillary for each leakage rate);
3. leakage increase valve and control
knob;
4. leakage decrease valve and control
knob;
5. vent (with protective cap) for
exhausting refrigerant-134a gas;
6. tank for holding liquid
refrigerant-134a (the tank contains
some refrigerant-134a when shipped
from the factory); and
7. a reservoir for holding refrigerant-134a
gas at a pressure corresponding to the
desired leakage rate.
Principles of Operation of
Adjustable Halogen Leak
Standard
The adjustable halogen leak standard
(Fig. 5) operates as discussed below. The
filler tank provides a supply of
refrigerant-134a liquid under its own
partial pressure. The increase valve
controls the amount of refrigerant gas fed
from the filler tank to the ballast tank, the
leakage rate meter and the leak capillary.
The pressure in the system is maintained
by the ballast tank. With the increase and
decrease valves closed, the system is
practically in a static state, except for the
minute amount of refrigerant gas that
escapes through the leak capillary. The
decrease valve provides a means of
decreasing the pressure built up in the
system. With the decrease valve opened,
refrigerant gas is allowed to escape
through the vent opening on the front of
the leak standard. The rate of refrigerant
gas escaping through the leak capillary is
a function of the pressure in the system
and is indicated on the leakage rate meter.
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
Preparation for Operation
of Adjustable Halogen
Leak Standard
The following procedure is used with
adjustable halogen leak standards.
1. Remove the protective caps from the
leak capillary in the probe fitting and
from the vent.
2. To increase the leakage rate, turn
increase valve knob counterclockwise
slowly until the instrument pointer
starts to move upscale. As the pointer
approaches the desired leakage rate,
gradually close the increase valve so
the pointer will stop at the desired
leakage rate. If the instrument pointer
continues to go upscale, this indicates
that the increase valve is not firmly
closed. Always make sure the increase
valve is closed firmly. (Avoid running
the instrument pointer off scale. This
can subject the instrument to as much
as 500 percent over pressure. Although
the unit can withstand the overload,
repeated abuse may damage it.)
3. To decrease the leakage rate, turn the
decrease valve knob counterclockwise
slowly until the instrument pointer
starts to move downscale. As the
pointer approaches the desired leakage
rate, gradually close the decrease valve
so that the pointer will stop at the
desired leakage rate. If the instrument
pointer continues to go downscale,
indication is that the decrease valve is
not firmly closed. Make sure the
decrease valve is closed firmly.
4. After increasing or decreasing the
leakage rate, be sure both valves are
closed by turning knobs clockwise.
5. After increasing or decreasing the
leakage rate and noting that the valves
are firmly closed, wait about 60 s for
the leakage rate to stabilize before
calibrating the leak detector. When the
leakage rate is being decreased,
refrigerant-134a gas is allowed to
escape through the vent to
atmosphere. During this operation it is
best to remove the leak standard from
the test area to avoid building up a
background of halogen vapor at the
test site. If this is not possible, attach
the vent tubing to the vent and
discharge the gas from the test area
through a window or other vent.
6. The leak standard is now ready for use.
Applications of Adjustable
Halogen Leak Standard
1. To check the operation and sensitivity
of the halogen leak detector. The
probe of the detector to be checked is
moved past the probe fitting of the
leak standard, which is set at the
maximum leakage rate allowable for
any single leak on the item being leak
tested. If an adequate signal is
obtained, the leak detector has
sufficient sensitivity (or more) to
detect this rate of leakage.
2. To determine size of leaks. If the leak
standard is set so that the leak detector
gives the same signal for the leak
standard as for the leak, the leak
standard then indicates the size of the
leak, only if 100 percent pure
refrigerant-134a is in the test system.
3. To extend the useful life of the sensing
element of the halogen leak detector.
Users frequently replace the detector’s
sensing element long before the end
of its useful life. A sensing element
can be used until it no longer
responds to the desired standard leak
setting. Additionally, the leak standard
permits use of the lowest possible
heater current to provide adequate
leak detector sensitivity. These
practices increase element life and
result in reduced maintenance and
lower replacement costs.
4. To simplify establishment of leakage
rate specifications. The leak standard
makes feasible the establishment of
leakage rate specifications and
provides a uniform standard for
calibrating leak detectors at each
location of product inspection.
5. To improve product quality. By
calibrating leak detectors with the leak
standard, it becomes possible to locate
and repair all significant leaks. This
ensures that products are
manufactured in accordance with
leakage specifications.
Precautions for Adjustable
Halogen Leak Standard
The following precautions should be
applied when using the adjustable
halogen leak of Fig. 4. Never allow any
grease or liquid to enter the leak capillary,
as it may plug the leak or alter its leakage
rate. When the leak standard is not in use,
it is recommended that the instrument
pointer be set up scale and that the
protective caps be placed over the vent
and leak capillary. This must be done to
prevent plugging of the capillary.
The adjustable halogen leak standard of
Fig. 4 may be used in several ways.
Calibrated Reference Leaks
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
83
Operational Procedure
When Pressurized System
Contains 100 Percent
Refrigerant
With the leak standard prepared for use,
proceed as follows.
1. Turn on the leak detector and let it
warm up for the time prescribed in the
applicable leak detector instruction
book. Set the leak detector in the same
mode of operation as that to be used
during leakage testing.
2. Place the probe squarely against the
probe fitting on the leak standard (see
Fig. 4) and observe the indicator
reading. Remove the leak detector
probe tip from the leak standard probe
fitting. When the leak detector reading
has settled to a stationary indication,
pass the tip of the leak detector probe
past the probe fitting on the leak
standard at a rate of about 25 mm·s–1
(1 in.·s–1). The tip of the probe should
just graze the front circular edge of the
probe fitting and pass across the
center of the probe fitting as shown in
Figs. 4 and 9.
3. Repeat the procedure of step 2 above,
reducing or increasing the sensitivity
setting of the leak detector each time,
until the leak detector signal is
adequate for the specified leakage rate.
The results of this test will indicate the
allowed probing speed and the safety
factor required and provide the
operator with a feeling for the
difference in indications between a
FIGURE 9. Technique for checking leak
testing sensitivity with sniffer probe tip
moving past the orifice of an adjustable leak
standard.
Tip of sniffer probe leak detector
84
Leak Testing
100 percent tracer gas probe intake
and the signal obtained during normal
probing procedure.
Interpretation of Unknown
Leakage Rate from
Comparable Standard Leak
A leak that gives the same leak signal as
the standard is the same size as that
indicated by the leak standard. A larger or
smaller signal indicates a larger or smaller
leak, respectively. If it is desired to
determine the size of any leak that is
located, adjust the leak standard in small
leakage rate increments (waiting about
60 s after each change) until the signal
caused by the leak standard is the same as
that caused by the leak. The leak standard
then indicates the size of the leak in
question, in terms of its leakage rate.
Operational Procedure
When Pressurized System
Contains Less than 100
Percent Refrigerant
Halogen leak standards can also be used
to calibrate a leak detector when the
system being checked contains less than
100 percent refrigerant-134a. For
applications using mixed gases in
pressurized components, the leak standard
may be used to calibrate a leak detector.
However, a leak from the vessel (such as a
tank, pipe or steam condenser) that
produces the same leak signal as does the
leak standard will have a total leakage rate
that is approximately inversely
proportional to the percentage of
refrigerant-134a tracer gas in the
enclosure. For example, suppose that,
with 10 percent refrigerant-134a in the
vessel, the leak standard indication is 30 g
per annum (1 oz·yr–1). The total leakage
rate is then 100/10 × 30 = 300 g·yr–1
(10 oz·yr–1).
Halogen leak standards are also used to
calibrate a leak detector when test systems
contain a halogen tracer gas other than
refrigerant-134a, such as refrigerant-22,
refrigerant-114 or refrigerant-11. The
leakage rates for these other tracer gases
may be read directly in standard cubic
centimeter per second. Leakage rates in
ounce per year can be obtained by
multiplying the readings in standard
cubic centimeter per second by 5.5 × 104.
Readings in pascal cubic meter per second
can be obtained by dividing the reading
in standard cubic centimeter per second
by a factor of 10.
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
Measuring Atmospheric
Contamination with
Adjustable Halogen Leak
Standard
To measure the amount of atmospheric
contamination with a heated anode
halogen vapor leak detector, the
equipment required includes an
adjustable halogen leak standard, a
halogen leak detector and a pure air
supply. The procedure recommended by
the leak detector manufacturer is as
follows.
1. In contaminated test areas, with the
leak detector operating at an air flow
of 4 cm3·s–1 (0.5 ft3·h–1), allow the leak
detector to breathe pure air for about
1 min, then allow the leak detector to
breathe air from the contaminated
area. If the leak detector gives a signal,
the area is contaminated. Note the
magnitude of the leak detector signal.
Do not adjust the sensitivity setting of
the leak detector between this
measurement and that which follows.
2. Move the leak detector and leak
standard to an area where there is no
atmospheric contamination. Adjust
the leak standard so that when the
leak detector sniffs the reference leak,
the leak signal is the same as when the
leak detector sniffed air in the
contaminated area. Note the leakage
rate shown on the dial of the leak
standard. This is a measure of the level
of atmospheric contamination with
halogen vapors in the original
contaminated area measured in step 1.
3. If it is desired to determine
(approximately) the degree of
contamination of the contaminated
area in parts per million (µL·L–1) of
halogen gas, the reading of the leak
standard from Step 2 in ounces per
year can be multiplied by 16. For
example, if the indication on the
reference leak is 1.5 oz·yr–1 the
contamination level is 1.5 × 16 =
24 µL·L–1. (For a leakage in grams per
year, divide the number by 1.8 to
arrive at the number of parts per
million.)
4. When using a leak detector that has its
own integral pure air supply, an
indication of the degree of
atmospheric contamination with
halogens can be obtained by holding a
finger over the probe tip for 30 s and
then switching the leak detector to
manual zero with the other hand. If
the leak detector reading is then
greater than the indication received
with leakage of the rejection level, the
halogen contamination of the air in
the test area is excessive.
Use of Calibrator for
Halogen Leak Standard
The calibrator is an accessory designed to
check the accuracy of calibrated leaks,
leak capsules and halogen leak standards.
Three models of the calibrator differ in
the bore size of the calibrated glass
capillary tube, which is the major
component of the calibrator. A small bore
capillary tube is used for leaks from
3 × 10–5 to 3 × 10–8 Pa·m3·s–1 (3 × 10–4 to
3 × 10–7 std cm3·s–1). A larger bore
capillary tube is used for leaks from
3 × 10–4 to 3 × 10–7 Pa·m3·s–1 (3 × 10–3 to
3 × 10–6 std cm3·s–1). Another model is
supplied with both capillary tubes.
The major component of the calibrator
is a calibrated glass capillary tube.
Accessories are provided allowing the
capillary tube to be connected to and
supported by the halogen standard leak
under test. To perform a calibration, the
calibrator is attached to the standard leak
through a vent valve. A colored indicating
liquid is inserted into the open end of the
capillary tube and the vent valve is
opened. The indicating liquid is drawn
into the capillary tube by applying a
slight suction to the plastic suction tube
connected to the vent valve. The vent
valve is then closed, retaining all the
escaping halogen vapor in the capillary
tube of the calibrator. By noting the
amount of movement of the indicating
liquid in the capillary tube for a specific
period of time, the magnitude of the leak
from the standard can be calculated and
compared with the reading of the leakage
rate gage on the standard.
Calibrated Reference Leaks
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
85
PART 3. Operation of Standard (Calibrated)
Helium Leaks
Functions of Calibrated
Helium Reference
Standard Leaks
Calibrated helium standard leaks are
essential when helium is used as the tracer
gas in leak testing for quantitative leakage
rate measurements. Calibration serves to
determine the user’s ability to detect
leakage and to perform quantitative
measurement of leakage rates. It is
imperative that the entire leak testing
system be calibrated. It is not sufficient
merely to calibrate the leak testing
instrument. In the case of a detector probe
test, for example, the detector probe must
be included in the leak testing system
during the calibration operations and used
in the normal manner as during testing.
The difficulty of repeating exact
detector probe techniques virtually
precludes the detector probe method as a
way of measuring leakage rates
quantitatively, although detector probe
tests are good qualitative tools. In vacuum
pumped systems, the system leak and the
artificial reference leak must be located
very close to each other for the
quantitative measurement to be valid.
Rating of Calibrated Helium Leaks
Calibrated helium leaks are usually
measured in units of pascal cubic meter
per second (Pa·m3·s–1) or standard cubic
centimeter per second (std cm3·s–1). It is
expected that standards in the future will
be calibrated in mass flow units of mole
per second. However, when discussing the
flow of a compressible fluid, it is necessary
to state not only the volumetric flow rate
(V/t) but also pressure P and temperature
T. Note that the units of leakage are
identical to the units of throughput (the
product PS of pressure P and pumping
speed S). Both leakage and throughput
describe the mass flow rate or, actually,
the number of gas molecules escaping per
unit time if the temperature is given.
Characteristics of Gaseous Flow
Involved in Leakage Calibrations
At least three additional variables must be
considered when using standard
calibrated leaks: (1) the nature of flow
(viscous, transitional or molecular) of gas
passing through the leak, (2) the specific
86
Leak Testing
tracer gas or gas mixture flowing through
the leak and (3) the pressure differential
acting across the leak.
In the viscous flow range, the mass
flow rate is inversely proportional to the
gas viscosity and directly proportional to
the difference in the squares of the
upstream and downstream pressures. In
the molecular flow range, the mass flow
rate is inversely proportional to the square
root of the mass of the gas molecule and
directly proportional to the difference in
partial pressure. Leakage at rates of
1 × 10–5 Pa·m3·s–1 (10–4 std cm3·s–1) or
greater will be most likely to be viscous
flow. Leakage at rates between 10–5 and
10–8 Pa·m3·s–1 (between 10–4 and
10–7 std cm3·s–1) will usually be
transitional in nature, exhibiting
characteristics of both molecular and
viscous flow. Leakage at rates in the range
of 10–8 Pa·m3·s–1 (10–7 std cm3·s–1) or
smaller will probably be molecular.
Membrane or Diffusion
Calibrated Reference
Helium Leaks
Two types of helium standard leaks or
calibrators are in general use, namely the
membrane type and the capillary type.
The design of a membrane or diffusion
type standard leak is shown in Fig. 2. This
standard leak has a reservoir filled with
helium surrounding a sealed glass tube
through which helium diffuses at a very
low rate, usually from 10–8 to
10–10 Pa·m3·s–1 (10–7 to 10–9 std cm3·s–1).
The standard calibrated helium leak
shown in Fig. 2 is fitted with a shutoff
valve that uses a metallic seal rather than
an elastomeric seal. This avoids spurious
changes in leakage rate due to helium
hangup. The reservoir is filled with
100 percent helium at 100 kPa (1 atm or
14.7 lbf·in.–2 absolute) of pressure. During
calibration of this leak, the pressure
differential feeding helium tracer gas into
an evacuated test system is therefore from
100 kPa to 0 kPa (14.7 to 0 lbf·in.–2).
Because the partial pressure of helium in
air is only about 0.5 Pa (4 mtorr), the glass
membrane calibrator of Fig. 2 continues
to leak helium even when it is not under
vacuum.
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
Capillary Standard
Calibrated Helium Leaks
(1)
The capillary helium leak consists merely
of flattened tubing or glass capillary
enclosed in a protective metallic sheath.
They are generally calibrated with one
end at vacuum and the other end at
atmospheric pressure. Capillary leak
standards are available with fixed leakage
rates varying from 10–3 to 10–6 Pa·m3·s–1
(10–2 to 10–5 std cm3·s–1). These capillary
leaks may have self-contained helium
reservoirs. They are more susceptible to
drastic changes in leakage rate caused by
clogging or by foreign agents such as dust
or condensation than are membrane leaks.
Computation of Molecular
Flow Leakage Rates for
Other Gases from Helium
Leakage Rates
Although helium is commonly used as the
tracer gas for mass spectrometer leak
testing, it is usually necessary to determine
the rate at which air would leak through a
similar discontinuity. In the molecular
flow range (and only in the molecular
flow range), the air leakage rate will be
about 35 percent of the helium leakage
rate through the same pressure differential.
When molecular flow occurs, the flow rate
for one gas can be compared to the flow
rate of any other gas by use of Eq. 1:
Q2
=
M1
M2
Q1
where Q1 is flow rate for gas 1 (any units
of leakage rate), Q 2 is flow rate for gas 2
(same units of leakage rate) and M1 is
molecular mass for gas 1 (relative atomic
mass).
Computation of Viscous
Flow Leakage Rates for
Other Gases from Helium
Leakage Rates
If the flow rate has been identified as
corresponding to viscous flow for one gas,
the viscous flow rate for any other gas can
be determined by use of Eq. 2:
(2)
Q2
=
n1
Q1
n2
where Q1 is flow rate for gas 1 (any units
of leakage rate), Q 2 is flow rate for gas 2
(same units of leakage rate), n1 is viscosity
of gas 1 (pascal second) and n2 is viscosity
of gas 2 (pascal second).
Table 1 lists the viscosities and
molecular masses of helium, argon and
neon inert tracer gases, air, nitrogen,
ammonia and other gases and vapors
commonly encountered in leak testing.
The values of the viscosities and
molecular masses from this table can be
used in Eqs. 1 and 2 to compute leakage
TABLE 1. Physical properties of certain gases and vapors.
Gas
Air
Ammonia
Argon
Carbon dioxide
Dichlorodifluoromethane
Dichloromethane
Helium
Hydrochloric acid
Hydrogen
Krypton
Methane
Neon
Nitrogen
Nitrous oxide
Oxygen
Refrigerant R-134a
Sulfur dioxide
Sulfur hexafluoride
Water vapor
Chemical
Symbols
NH3
Ar
CO2
CCl2F2
CH2Cl2
He
HCI
H2
Kr
CH4
Ne
N2
N2O
O2
C2H2F4
SO2
SF6
H2O
Relative
Molecular
Mass (Mr)
29.00
17.03
40
44.01
120.93
84.83
4.00
36.50
2.02
83.80
16.04
20.18
28.01
44.00
31.99
102.03
64.00
146
18.02
Gas
Constant
(J·kg–1·K–1)
287
488.22
207.86
188.89
68.75
98
2078.60
227.79
4116.04
99.22
518.35
412.01
296.84
188.96
259.91
81
129.91
57
461.40
Viscosity at 15 °C (59 °F)
________________________
µPa·s
(millipoise)
174
97
220
145
127
(1.74)
(0.97)
(2.20)
(1.45)
(1.27)
192
140
86
246
107
309
173
143
199
(1.92)
(1.40)
(0.86)
(2.46)
(1.08)
(3.09)
(1.73)
(1.43)
(1.99)
123
152
93
(1.23)
(1.52)
(0.93)
Calibrated Reference Leaks
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
87
rates for other gases from helium leakage
rates determined by helium leak tests.
Computation of
Transitional Flow Rates
Transitional flow, flow between the
viscous and molecular regimes, is not
fully understood. A number of models
have been proposed to estimate the flow
in this regime. The simplest and most
used is the slip flow model. The flow is
given by the following equation,
(3)
Q2
πr 4
=
−
×
(P
8n l
1
Pa

λ
− P2  1 + 4 
r 

)
where r is tube radius (meter), l is tube
length (meter), P1 is pressure upstream
(pascal), P2 is pressure downstream
(pascal), λ is mean free path and Pa = (P1
+ P2)/2. If the tube diameter or effective
diameter is known, the flow for any gas
can be calculated with the equation from
the known pressures.
Also, the flow for gas can be estimated
from the known flow for another gas
under the same pressure conditions.
Alternately, a crude estimate could be
made using the formula for molecular
flow. In general, estimation of flow based
on the known flow of another gas in the
transition range is not recommended.
Effect of Absolute Gas
Temperature on Molecular
Flow Leakage Rates
The effect of absolute gas temperature on
conductance when the leakage flow is
molecular should not be overlooked when
estimating leakage rates by use of
standard leaks. The conductance of both
orifices and of tubes changes directly with
the absolute gas temperature. Equation 4
shows how the new flow rate Q2 at the
new absolute temperature T2 (K) compares
with the original flow rate Q1 at absolute
temperature 1, through a leak for which
both the leak path dimensions and the
pressure difference across the leak remain
constant, for the specific case of the
molecular flow rate at new absolute
temperature:
(4)
88
Leak Testing
Q2
=
T2
Q1
T1
Varying the Pressure
Differential across a Leak
Because the rate of flow of a gas through a
leak will be a function of either the
molecular mass or the viscosity of the gas
flowing, it is sometimes very important to
know which type of leakage flow is
occurring. This is especially true if a
leakage rate must be expressed in terms of
one gas such as air, when leakage must be
measured by detecting helium flow.
Often, the capillary leak is used under
conditions that vary greatly from the
conditions under which the leak was
calibrated. The test gas, test pressure or
both may be different from those used in
the calibration of the leak. If it is not
possible to obtain a true calibration figure
under the new test conditions, it becomes
necessary to attempt an estimation. The
principles used in such estimations are
presented next.
Varying System Pressure to
Identify Types of Flow in
Leaks
If leakage, or flow, can be measured by
using a leak detector or a residual gas
analyzer, the type of flow can often be
identified by changing the pressure
causing the flow of gas. All techniques of
leak testing using a mass spectrometer
leak detector involve the passage of a
tracer gas through a presumed leak in a
pressure barrier. This involves application
of tracer gas to the high pressure side of
the barrier and the subsequent detection
of the tracer gas on the lower pressure
side. In general, there are three types of gas
flow: viscous, transition and molecular.
The variables that control the type of gas
flow that occurs in leaks are (1) viscosity
of the flowing gas or gas mixture (Pa·s),
(2) relative molecular mass Mr of the gas,
(3) pressure difference causing the flow
(Pa), (4) absolute pressure in the system
(Pa absolute) and (5) absolute temperature
of the flowing gas or gas mixture (K).
Figure 10 shows the general relationship
of flow type to gas pressure and radius of
tubular conductance.
Conditions for Identification of
Viscous Flow through Leaks
When the pressure across a leak is
changed and the flow changes in
proportion to the differences of the
squares of the absolute pressures, the
leakage can be identified as viscous flow.
Viscous flow occurs in high pressure
systems, such as systems pressurized with
helium tracer gas and checked by the
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
helium detector probe method. Figure 11
shows graphically how the viscous leakage
rate changes as internal pressure is varied
in test objects and systems leaking to the
atmosphere. Figure 12 shows similar
graphical relationships for externally
pressurized test objects with leakage to an
internal volume that is highly evacuated.
Conditions for Identification of
Molecular Flow through Leaks
If the flow of gas through a leak changes
in proportion with the difference between
the pressures acting across the leak, the
flow of gas is molecular. Molecular flow
usually occurs in vacuum testing
applications with helium spray
application of tracer gas and mass
spectrometer leak detectors attached to
the internal volume of evacuated test
objects. Figure 13 shows graphically how
the molecular leakage rate varies linearly
with the pressure differential as external
pressure is varied on test objects and
systems that are internally evacuated.
Conditions for Identification of
Transitional Flow through Leaks
If the flow changes in response to changes
of pressure by some relation between
proportionality to differences in squares
of pressures and proportionality to
difference in pressures, the leak involves
transitional flow. Figure 10 illustrates the
regimes for each of these three types of
105
(4 × 103)
104
(4 × 102)
103
(4 × 101)
102
(4 × 100)
101
(4 ×
100
(4 × 10–2)
Calculating Effect of
Pressure Changes with
Viscous Flow through
Leaks
Viscous flow occurs when the mean free
path of molecules of the gas is much
smaller than the cross sectional
dimension of the physical leak. In this
case, the leakage rate Q is proportional to
the differences in the squares of the
pressures on the opposite sides of the
pressure barrier through which the leak
penetrates. If the viscous flow rate Q1 has
been determined for a difference between
pressure P1 and pressure P2 and then the
pressures are changed to new values P'1
and P'2, the new flow rate Q 2 can be
calculated by means of Eq. 5, for viscous
flow rate at a new pressure,
(5)
Q2
=
P ′12 − P ′22
P12 − P22
Q1
FIGURE 11. Graphical presentation of increase in viscous flow
leakage ratio when pressurizing with 100 percent tracer gas,
as a function of internal system pressure when leaking to
atmospheric air.
100
Leakage rate increase ratio
Radius of tube, mm (in.)
FIGURE 10. Graphical presentation of conditions for viscous,
molecular and transitional flow of gases through leaks, in
terms of absolute gas pressure at 25 °C (77 °F) and radius of
tubular conductance. Note that 1 Pa = 1.5 × 10–4 lbf·in.–2.
flow of gases through leaks as a function
of absolute gas pressures and diameter of
leak passageways. In many cases, it is not
always practical to vary pressures on parts
under test to determine the types of leaks
being detected. In instances where the
leakage of a gas other than the tracer gas
is of concern, it is best to assume the
worst possible condition, which may be
either viscous or molecular flow.
Viscous
10–1)
Transition
10–1 (4 × 10–3)
10–2 (4 × 10–4)
Molecular
10–3 (4 × 10–5)
100 000
Left
scale
10
10 000
Viscous flow
Right
scale
1.0
1000
10–4 (4 × 10–6)
10–5 (4 × 10–7)
mPa
kPa
Pa
10–3 10–2 10–1 100
101
102
Pressure (Pa)
103 104
MPa
105 106
100
100
(15)
1000
(150)
10000
(1500)
100 000
(15 000)
Absolute pressure, kPa (lb f ·in.–2)
(outside of part at 100 kPa)
Calibrated Reference Leaks
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
89
FIGURE 12. Graphical presentation of increase in viscous flow leakage rate ratio when
pressurizing a chamber with 100 percent tracer gas, as a function of external system pressure,
when leaking into internally evacuated test objects in the chamber.
1 000 000
1000
800
600
500 000
400
Leakage rate increase ratio
200
Left scale
100 000
100
80
60
50 000
40
Viscous flow
20
10
8
6
10 000
Right scale
5000
4
2
1000
1
0.10
(0.015)
0.30
(0.045)
1.00
(0.150)
3.00
(0.45)
10.0
(1.50)
30.0
(4.50)
100.0
(15.0)
Absolute pressure, MPa (lb f ·in.–2 × 103)
(inside of part at high vacuum)
FIGURE 13. Graphical presentation of increase in molecular flow leakage rate ratio with
molecular flow, as a function of external pressure of 100 percent tracer gas, when leaking into
internally evacuated test objects of systems (inside of test system or parts at high vacuum).
1000
800
600
400
Molecular flow
Leakage rate increase ratio
200
100
80
60
40
20
10
8
6
4
2
1
0.10
(0.015)
0.30
(0.045)
1.00
(0.150)
3.00
(0.45)
10.0
(1.50)
30.0
(4.50)
100.0
(15.0)
External pressure, MPa (lb f ·in.–2 × 103) absolute
(inside of part at high vacuum)
90
Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
In Eq. 5, the pressures are all absolute
pressures in pascal or pound per square
inch (lbf·in.–2). The old and new flow rates
must be in the same units of leakage.
Equation 5 for viscous flow through leaks
would apply for leak testing of systems at
higher than atmospheric pressure. It
applies to a helium detector probe test on
an internally pressurized test system
leaking to the atmosphere.
Example Calculation of Capillary
Leakage Rate at Different
Pressures
Assume that a capillary standard leak
(flattened tube) has been calibrated for a
nitrogen flow rate of 2 × 10–5 Pa·m3·s–1
(2 × 10–4 std cm3·s–1) with atmospheric
pressure on the high side and zero
pressure (vacuum) on the low side. It is
desirable to predetermine the leakage rate
if this same capillary leak is to be used
with twice atmospheric pressure on the
high side and atmospheric pressure on the
low side. (Note that the pressure
differential between high and low sides of
the leak is atmospheric pressure of
100 kPa (1 atm) in both old and new
cases.) Because the leakage rate is so high,
it will be assumed that leakage occurs as
viscous flow. By Eq. 5, the new flow rate
Q2 is calculated in Pa·m3·s–1:
Q2
=
2 2 − 12
12 − 0 2
=
6 × 10 − 5
2 × 10 − 5
This new flow rate represents a threefold
increase when compared to the original
flow rate obtained with internal
atmospheric pressure leaking to vacuum.
Example Calculation of Leakage
Rate after Pressuring Up Helium
with Nitrogen
Another situation often encountered in
mass spectrometer leak testing involves a
standard capillary leak used with a
mixture of helium and nitrogen at high
pressure. This case occurs most commonly
when the user attempts to increase
helium leak testing sensitivity by the
technique of pressuring up. This
technique is used, for example, when a
large volume test object or system is
tested with helium tracer gas and leaks are
detected with a detector probe. The vessel
under test is originally filled with air at
atmospheric pressure. The calibrated
capillary leak is attached to the vessel and
absolute pressure is raised to a total of
200 kPa (2 atm) by injection of helium
tracer gas. Then compressed air or
nitrogen is forced into the vessel, raising
its absolute pressure even higher, for
example, to 400 kPa (4 atm). (For large
test volumes, 100 percent helium at high
pressure may not be economical.)
In this example of pressuring up, the
new total viscous flow rate Q2 can be
estimated:
Q2
4 2 − 12
=
2 × 10 − 5
12 − 0 2
30 × 10 − 5
=
The actual helium leakage rate, because
the final pressurized mixture is only
25 percent helium, is only about
7 × 10–5 Pa·m3·s–1 (7 × 10–4 std cm3·s–1) of
helium. The result of pressuring up with
air or nitrogen is an approximately linear
increase in the helium flow rate through
the leak. This example calculation would
be valid only for viscous leakage. (Note
that 1 Pa·m3·s–1 = 10 std cm3·s–1.)
Limitations of Increasing Pressure
with Molecular or Transitional
Flow Leaks
For molecular flow leaks, increasing
pressure with air would not result in an
increase in the helium flow rate. In the
transitional flow range, particularly when
dealing with gas mixture, the situation
(degree of enhancement of leak signals) is
extremely difficult to predict. In these
cases of unknown effects, it would be
useful to make a graphical plot of leak
signal amplitude as a function of total
pressure within the leaking vessel to aid
in determining the nature of a leak.
Correction for Aging of
Helium Membrane
Calibrated Leaks
As time passes, the internal helium
pressure of glass permeation leaks is
depleted (see Fig. 14). This depletion
results from the gas leaking from the
reservoir through the glass element and
through any discontinuities in the
reservoir into the atmosphere. If there is
no appreciable leakage except through the
glass permeation membrane, then the
depletion rate of the leak can be estimated
from the original number of moles of
helium in the reservoir and the original
leakage rate as follows. The leak rate N(t)
at a time t after the original calibration
N(to) can be determined from
(6)
N (t )
=
N (t o ) e
−
N (t o )
C (t o )
⋅ t
where the leakage rates are in mole per
second, C(to) is the original number of
moles of gas in the reservoir and t is the
elapsed time since the original calibration.
Calibrated Reference Leaks
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
91
The amount that the leakage rate changes
as a function of time depends on the
design of the calibrated leak and on usage
conditions and can vary from less than
1 percent per year to more than
20 percent per year.
Correction for
Temperature of Helium
Membrane Standard Leaks
The permeation rate of helium through
glass is described by an exponential
expression:
(7)
Q
=
AT e
−
b
T
where A (Pa·m3·s–1·K–1) and b (K) are
constants and T is the absolute
temperature (K). Table 2 gives typical
values of the temperature coefficient b.3
TABLE 2. Temperature coefficients (measured by the
National Institute of Standards and Technology) and
corresponding glass types for helium permeation leaks.
Temperature
Coefficient (K)
Probable
Glass Type
≤ 2500
2700
3000
3600
borosilicate
fused silica
Pyrex® 7740
Corning® 7052
FIGURE 14. Decline in leakage rate as a function of depletion
rate.
Percent loss in leakage rate
100
10
1
0.1
0.1
1
Frequently a linear approximation is
used:
(8)
Q
=
[
Q c 1 + a (T − Tc
)]
where Qc is the leak rate at the calibration
temperature, Tc is the calibration
temperature, T is the temperature at test
conditions and a is a linear temperature
coefficient, about 0.03 °C–1 (0.05 °F–1).
Using the linear temperature expression
will generally give adequate
representation over small temperature
variations, less than 5 °C (9 °F). For
temperatures differences of 30 °C (54 °F),
errors as large as 75 percent can be made
using this simplifying assumption. For the
lowest uncertainties the leak should be
calibrated close to the temperature at
which it will be used. A rough
approximation to the linear correction is
given in Fig. 15.
Glass Capillary Leaks for Tracer
Gases Other than Helium
For gases other than helium, such as
argon, neon or hydrogen, permeability
rates in glass become small. The most
common calibrating leak for these gases is
a glass capillary leak (glass being chosen
for its ease of fabrication of small capillary
tubes and orifices). There are two areas in
which these glass capillary leaks differ in
characteristics from glass membrane leaks.
1. Depletion of internal pressure due to
aging of capillary glass leaks is a
function of the length of time the
standard leak is in use, because the
rate of gas flow in a capillary leak is a
function of the total pressure drop
across the reference leak (not the
helium partial pressure difference that
controls the flow rates of helium
through glass membrane leaks). In a
capillary leak, there is no helium flow
unless the leak is being pumped on a
vacuum pump or leak detector system.
2. Glass capillary standard leaks exhibit a
negative temperature coefficient. This
means that the capillary tube or orifice
must decrease in diameter as
temperature rises. This diameter
reduction reduces the gas flow at a
faster rate than the internal pressure
rise increases the flow rate as
temperature rises.
10
Annual leak depletion rate for leak
(percent for year)
Legend
= 4 years
= 2 years
= 1 year
92
Leak Testing
= 6 months
= 3 months
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
FIGURE 15. Temperature correction factor for a silica membrane standard helium leak used at
operating temperatures that differ from temperature during initial calibration. To correct
calibrated leakage rate for temperature, multiply by correction factor.
3.0
2.0
1.9
1.8
1.7
1.6
1.5
1.4
1.3
1.2
eit
nh
re
ah
F
Ce
lsi
us
Correction factor
2.5
1.1
1.0
0.9
0.8
0.7
0.6
0.5
0.4
–50
(–90)
–40
(–62)
–30
(–54)
–20
(–36)
–10
(–18)
0
10
(18)
20
(36)
30
(54)
40
(72)
50
(90)
Temperature difference, °C (°F)
Calibrated Reference Leaks
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
93
PART 4. Calibration of Standard Reference
Leaks
Commercial Sources for
Calibrated Leaks
Commercially available permeation leaks
have been limited to helium in the past
because glass elements were
predominantly used. It is now possible to
obtain polymer permeation elements that
function with other gases including
argon, sulfur hexafluoride and many
refrigerants. In addition many calibrated
physical leaks are also commercially
available. The choice of gases in these
physical leaks, predominantly capillary
type, are large and include most
noncorrosive, nontoxic industrial gases.
Calibration Techniques for
Artificial Physical
Reference Leaks
Calibrated leaks are available with eight
decades of leakage values. Because of this
large range of leakage, calibration is
difficult. The five techniques of measuring
leakage rates are (1) isobaric volume
change, (2) pressure rise, (3) pressure
drop across a known conductance,
(4) pressure measurement at constant
pumping speed and (5) comparison.
Isobaric Volume Change
Calibration of Standard Leaks
A schematic diagram of the equipment
used in the isobaric volume change
technique of leakage rate measurement is
shown in Fig. 16. One side of the leak is
attached to a vacuum system; the other
side is attached to a gas reservoir at
atmospheric pressure. To this reservoir is
attached a capillary of known cross
section. A slug of indicating fluid is placed
in this capillary. As gas leaks from the
volume into the vacuum, the slug of fluid
travels down this capillary to keep the
pressure in the reservoir constant. The
leakage rate is determined by
measurement of the volume displaced by
the travel of the slug down the capillary:
(9)
Q
=
P (V 2 − V1 )
t
where Q is leakage rate (Pa·m3·s–1), P is
pressure in the gas volume (pascal),
V2 – V1 is volume displaced during travel
of the indicating fluid (cubic meter) and t
is time (second).
Limitations of Isobaric Volume
Change Leak Calibration
The primary limitation of the isobaric
volume change technique is the size of
the capillary tube involved in the volume
measurement (see Fig. 16a). It is difficult
to obtain a liquid that can be placed in a
small capillary tube and that subsequently
can be forced out the other end. For this
reason, the practical limitation of the
capillary tube technique of volume
displacement measurement is in the range
of 1 × 10–6 m3·s–1 (2 × 10–3 ft3·min–1). It
would theoretically be possible to use a
slightly larger capillary and to take longer
periods of time between readings but
errors might arise from permeation of gas
either through the liquid slug or through
FIGURE 16. Leak calibration by isobaric volume change:
(a) with capillary tube; (b) with differential pressure gage.
(a)
Gas at atmospheric pressure
Graduated capillary
Slug of liquid
indicator
Vacuum
Leak
Vacuum pump
(b)
Differential
pressure gage
Piston
P
Vacuum
Leak
Gas at
atmospheric
pressure
Vacuum pump
94
Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
the walls. An error might also be
introduced by a change in barometric
pressure or a change of ambient
temperature. This becomes particularly
critical in the calibration of small leaks,
because a slight temperature change
might produce a volume change greater
than that due to efflux of gas.
Selection of Liquid for Capillary
Slug That Indicates Volume
Change
It is desirable that the indicating fluid not
be permeable to the gas being calibrated.
For this reason, mercury is the preferred
indicating fluid. Because of its high
surface tension, mercury cannot be placed
in a small capillary. This drastically limits
the size of the leak that may be calibrated
with mercury. For these measurements it
is desirable to use a liquid with a low
vapor pressure so that the leak is not
contaminated by the calibration fluid.
Unfortunately, most liquids of low vapor
pressure are also of high viscosity and
make it difficult to obtain an accurate
measurement of the flow of liquid
displaced in the capillary. These fluids also
tend to form bubbles at the end of the
capillary. The added pressure necessary to
remove the bubble of liquid from the end
of the capillary prevents the experiment
from being isobaric.
(10) Q
= V
dP
dt
where Q is leakage rate (Pa·m3·s–1), V is
volume of evacuated chamber (m3), P is
pressure in chamber (Pa absolute), t is
time (s) and dP/dt is time rate of pressure
change (Pa·s–1).
Limitations of Leak Calibration by
Pressure Rise
The major difficulties with the pressure
rise calibration technique occur in
measurement of pressure. The pressure in
an evacuated system usually does not stay
constant, but gradually increases due to
outgassing of the walls of the chamber.
The pressure rise due to this desorption
must be taken into account in
calculations. The type of pressure
instrumentation to be used for the
measurement depends on the range of
pressures that are expected to be
measured. Table 3 lists some gages that
may be used and their ranges.
The effect of desorption on the
uncertainty of the measurement will
depend on the ratio of the apparent
leakage because of gas desorption to that
of the leakage to be measured. It should
be recognized that the rate of desorption
is usually not constant and will in general
be a function of temperature.
Piston and Differential Pressure
Gage in Isobaric Volume Change
Tests
Another technique for measuring volume
displacement is with a piston to replace
the effluent gas. In this technique, a
differential pressure gage is used to
measure the pressure in the gas volume
and the piston is manually pushed into
the volume at such a rate as to keep the
pressure constant. The pressure gage need
not be calibrated because the readings are
made only when the differential pressure
gage is at zero indication. This technique
can readily measure leakage as low as
10–9 Pa·m3·s–1 (10–8 std cm3·s–1).
TABLE 3. Gages for pressure rise leak calibration.
Pressure Range
_________________________
Gage
(lbf·in.–2)
Pa
Mass spectrometer
< 10–3
Molecular drag
10–4 to 10–10
Capacitance diaphragm
10–1 to 105
(< 1.5 × 10–7)
(1.5 × 10–8 to 1.5 × 10–14)
(1.5 × 10–5 to 1.5 × 101)
FIGURE 17. Leak calibration by pressure rise technique.
Pressure Rise Calibration
of Standard Leaks
The second technique of calibrating leaks
is by means of the pressure rise technique.
A leak and its gas supply are attached to
an evacuated chamber of known volume
in the arrangement sketched in Fig. 17.
The leaking gas is allowed to accumulate
in this volume and the pressure rise is
measured at various intervals. The leakage
may then be computed by Eq. 10:
Gas at
atmospheric
pressure
Pressure
gage
Leak
Vacuum
P
Vacuum pump
Calibrated Reference Leaks
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
95
Practical Example of Leak
Calibration by Pressure Rise
As noted just previously, the pressure rise
technique for calibrating a standard
reference leak in a laboratory depends on
the outgassing surface area as well as the
volume of the system to be evacuated.
The upper size limit for large leaks to be
measured by this technique would be
governed mainly by the largest size of test
volume that could be realistically placed
within a laboratory. Probably leaks as
large as those with 1 Pa·m3·s–1
(10 std cm3·s–1) leakage rates would be
near the upper limit.
The size limit for small leaks measured
by the pressure rise calibration technique
would be governed by the accuracy of
measurement of the volume of the
evacuated test system and the accuracy
with which the pressure change could be
measured. These may place the lower
limit of leak size in the range of leakage
rates from 10–5 to 10–6 Pa·m3·s–1 (10–4 to
10–5 std cm3·s–1).
In a practical industrial laboratory,
calibration would be performed by
measuring the rate of pressure rise of a
well conditioned evacuated system volume
when closed off from external sources of
gases. The result may be a curve similar to
the lower curve shown in Fig. 18.
Following that test, the reference
physical standard leak would be attached
to the same test volume, which would be
evacuated to the same vacuum level as in
the first test, with the valve closed
between the chamber and the standard
FIGURE 18. Pressure rise as a function of time elapsed after
evacuating test chamber, during calibration tests of physical
reference leak.
leak. Again, the pressure rise of the
evacuated system would be measured, this
time with the valve open so that air enters
through the standard leak to be
calibrated. The rate of pressure rise is
higher with the leak in place and the
curve from the second test with the leak
admitting air to the evacuated chamber
would be higher than the initial curve, as
indicated by the higher curve of Fig. 18.
The vertical difference between the two
curves (with and without the leak opened
to the evacuated chamber) indicates the
theoretical rate of rise or pressure due to
the leak.
With this number, together with the
values for system volume and test time,
the rate of leakage through the standard
leak under calibration test can be
calculated. Figure 19 shows the relation of
pressure difference to the elapsed test time
and approaches a linear (straight line)
relationship. The leakage rate is computed
in SI units from the relation
(11) Q
= V
∆P
t
where V is volume (cubic meter), ∆P is
pressure difference (pascal) and t is
elapsed test time (second). For example,
for the case shown in Fig. 19, the
calculation is as follows:
Q
0.0169
700
=
0.382 ×
=
9.2 × 10 −6 Pa ⋅ m 3 ⋅ s −1
=
9.2 × 10 −5 std cm 3 ⋅ s −1
FIGURE 19. Pressure difference resulting from leakage
through standard leak.
20.0 (150)
18.7 (140)
17.3 (130
Pressure rise with leak attached
Air
Pressure, mPa (µtorr)
13.33 (100)
Pressure rise without leak
1.33 (10)
0
100
200
300
400
500
600
700
Pressure, mPa (µtorr)
16.0 (120)
14.7 (110)
13.3 (100)
12.0
(90)
10.7
(80)
9.3
(70)
8.0
(60)
6.7
(50)
5.3
(40)
4.0
(30)
2.7
(20)
1.3
(10)
0
0
0
Pressure rise elapsed time (s)
96
Leak Testing
100
200
300
400
500
600
700
Pressure rise elapsed time (s)
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
Calibration of Standard
Leaks by Pressure Drop
across a Known
Conductance
Calibration of Standard
Leaks by Pressure
Measurement at Constant
Pump Speed
A third technique of measuring leakage
rates is by measuring the pressure drop
across a known conductance C. This
technique is illustrated in Fig. 20. The
pressure drop (P1 – P2) across a known
conductance is proportional to the flow
rate Q, as indicated by Eq. 12:
A fourth technique of calibrating the flow
of a leak is by measuring the pressure it
produces in a vacuum system that is
pumped at a known speed (see Fig. 21).
This is the limiting case for Eq. 12, when
P2 is zero. The equation then being used
takes the form of Eq. 13:
(12) Q
=
C ( P1 − P2 )
(13) Q
With molecular flow the conductance
C may be designed from theoretical
grounds and such a conductance can be
accurately constructed.
Limitations of Pressure Drop Leak
Calibration Technique
The major difficulty with the pressure
drop calibration technique is in obtaining
accurate pressure measurements.
Ionization gages have been used for the
pressure measurement in evacuated
systems, but their readings are often
questionable. Because their sensitivities
are more often in doubt, pressure drop
leakage tests are also used to calibrate
ionization gages.
An alternative to using an ionization
gage is to use a molecular drag gage,
sometimes referred to as a spinning rotor
gage. This instrument is stable with time
and can achieve accuracies of ±10 percent
even if uncalibrated over the pressure
range of 10–4 Pa to 10–1 Pa (1.5 × 10–8 to
1.5 × 10–5 lbf·in.–2).
FIGURE 20. Leak calibration by pressure drop across a known
conductance.
Pressure
gages
Gas at
atmospheric
pressure
P1
=
SP
where S is the pumping speed (m3·s–1) of
the system (usually governed by an
orifice) and P is the ultimate pressure
(pascal) attained within the vacuum
chamber while being pumped. The system
is usually constructed so that the
pumping speed is controlled by molecular
kinetics considerations and can be
rigorously calculated from theoretical
grounds. The disadvantage of the
pumping speed technique is, again, that
the pressure of the system must be
accurately measured. If the leakage Q and
the pumping speed S are known, P can be
derived using the above equation. This
type of system has also been used to
calibrate pressure gages.
Mass Spectrometer As Pressure
Gage in Leak Calibrations
If a mass spectrometer is used as the
pressure gage, some accuracy is gained
because the error due to outgassing is
minimized. The pumping speed system
has essentially the same flow pattern as
the mass spectrometer leak detector. In a
mass spectrometer leak detector, Eq. 13
takes the form of Eq. 14:
FIGURE 21. Leak calibration by pressure measurement at
constant pumping speed.
Pressure
gage
Gas at
atmospheric
pressure
P2
Leak
Leak
P
Limiting
conductance
Conductance
Vacuum pump
Vacuum pump
Calibrated Reference Leaks
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
97
(14) Q
=
S Ka
=
C1 S
where Q is leakage rate (Pa·m3·s–1); S is
pumping speed, a constant (m3·s–1); K is
conversion factor for pressure from
collector current reading (Pa·mA–1); a is
collector current reading in mass
spectrometer (mA); and proportionality
constant C1 equals Ka (Pa). In most cases,
a and K are not known, but a
proportionality constant C1, the product
of these two numbers, is used. Providing
that the response of the leak detector is
linear, the mass spectrometer can be used
to calibrate leaks by comparison to
standards calibrated by other techniques.
Comparison Calibration
A fifth type of calibration is by
comparison with a calibrated leak whose
measurement is traceable to the National
Institute of Standards and Technology.
This technique can be used over a wide
range of leakage rates, 10–11 to
10–3 Pa·m3·s–1 (10–10 to 10–2 std cm3·s–1)
and with a wide range of gases. With a
mass spectrometer type leak detector
(helium only) a calibrated leak may be
compared to a leak whose leakage rate is
to be determined. The leakage rate is
calculated with the following expression:
(15) Q unk
=
Q std
H unk
H std
where Qunk is the leakage rate of the
unknown leak and Qstd is the leakage rate
of the calibrated leak, H is the mass
spectrometer signal corresponding to the
two measured leakage rates. This equation
assumes that the mass spectrometer gives
a linear response to partial pressure
changes and that the pumping speed of
the system is stable over the testing
period. To minimize uncertainties due to
nonlinearities in the mass spectrometer
the calibrated leak should be closely
matched to the unknown leak. For the
most accurate results, it is usually
necessary to have the two leaks register
signals in the same decade of the
measuring instrument.
Matching Standard
Leakage Rate to
Permissible Leakage Rates
The standard or reference leakage rate
used in leak testing should be of the
approximate value of the permissible
leakage rate of the test object. This must
be so if the response of the detector to
leakage is not linear. The smaller the
standard leakage rate, the greater the
difficulties associated with it. If the
standard leak is substantially different
from the permissible leakage (a
contingency that may result from the
difficulty of making small standard leaks),
the response of the detector to different
leakage rates becomes important.
Calibration of Standard
Leaks with Different Gases
Basic leakage rate measurements are
necessary for the calibration of primary
standard leaks used in connection with
tracer gas leakage rate measurement
systems. Fortunately, controlled
laboratory conditions are practical for
such calibrations and time is not an
essential factor. Figure 22 shows
schematically two basic systems, constant
FIGURE 22. Standard leak calibration, dQ/dt: (a) pressure change calibration system;
(b) volume change calibration system.
Pressure gage
(McLeod)
(a)
Gas supply or vacuum
}
V
{
Vacuum system
{
Vacuum system
Leak
(b)
Pressure gage
Gas supply or vacuum
}
Leak
V
98
Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
volume and constant pressure. The
leakage rate Q for the pressure change
calibration system of Fig. 22a is given by
Eq. 16:
(16) Q
=
d ( PV )
dt
= V
dP
dt
The leakage rate Q for the volume
change calibration system of Fig. 22b is
given by Eq. 17:
(17) Q
=
d ( PV )
dt
=
P
Closing
Calibrated leaks have a vital role in leak
testing programs. Attention to the proper
calibration techniques can enhance the
operator’s understanding of the test
procedure and hence improve the
reliability of the leak testing being
performed.
dV
dt
It may be noted that no reservoir is
shown for the leak in Fig. 22. Elimination
of a fixed upstream leak reservoir has two
important advantages. First, using a
vacuum upstream permits an evaluation
of outgassing and other extraneous
sources of gas arising in the calibration
system. Second, the same leak element
can be calibrated for many leakage rate
values for various gases simply by varying
the upstream gas and pressure.
For leaks in the 10–9 Pa·m3·s–1
(10–8 std cm3·s–1) range, accumulation
times as long as a week have been used
for increasing measured pressure change
in the constant volume manifold V of the
V(dP/dt) calibration system. Conversely,
times of the order of 100 s have been used
for increasing the pressure in the known
volume V from an insignificant pressure
to an arbitrary pressure of 0.4 Pa
(5.8 × 10–5 lbf·in.–2) in the P(dV/dt) system
for leaks in the 10–5 Pa·m3·s–1 range.
The procedure for this second
technique is as follows. The known
volume V is evacuated to a negligible
pressure, for instance less than 10 mPa
(1.5 × 10–6 lbf·in.–2), and then valved off.
The valve to the vacuum system is then
closed; thereafter, gas from the leak is
admitted to the manifold. At the instant
the pressure in the manifold attains a
preselected value P, a timer is started.
Opening the valve to V will lower the
manifold pressure temporarily, but the
pressure will again increase steadily
because of the continued inflow of gas
from the leak. When the pressure again
climbs to the value P, the time is stopped.
The only difference between the
conditions when starting and stopping
the time is that the pressure in V
increased from essentially O to P.
Calibrated Reference Leaks
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
99
References
1. Nondestructive Testing Handbook,
second edition: Vol. 1, Leak Testing.
Columbus, OH: American Society for
Nondestructive Testing (1982).
2. Ehrlich, C.D. and J.A. Basford.
“Recommended Practices for the
Calibration and Use of Leaks.” Journal
of Vacuum Science and Technology A —
Vacuum, Surfaces, and Finishes. Vol. 10,
No. 1. New York, NY: American
Institute of Physics, American Vacuum
Society (Jan.-Feb. 1992): p 1-17.
3. Abbott, P.J. and S.A. Tison.
“Commercial Helium Permeation Leak
Standards: Their Properties and
Reliability.” Journal of the Vacuum
Society of America A — Vacuum,
Surfaces, and Finishes. New York, NY:
American Institute of Physics,
American Vacuum Society (May-June
1996): p 1242-1246.
100
Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
C
4
H A P T E R
Safety Aspects of Leak
Testing
Gerald L. Anderson, American Gas and Chemical
Company, Northvale, New Jersey
Robert W. Loveless, Nutley, New Jersey
Charles N. Sherlock, Willis, Texas
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
PART 1. General Safety Procedures for Test
Personnel
Test Personnel Dedication
to Safety Procedures
The range of applications of leak testing is
so wide and varied that no single set of
safety rules for protection of personnel
and property can be made to cover all
cases. Leak testing personnel must be
made aware of job hazards and be
receptive to proper training to protect
themselves and others working close by.
On many jobs, testing must be performed
at odd hours and under awkward
conditions. Nightshift work, weekend
work and work in unheated areas in
winter and uncooled areas in summer are
common. Climbing through manholes,
climbing ladders and scaffolds, balancing
on structural members or other awkward
maneuvers may all be in a day’s work.
In addition to technical abilities and
training in test procedures, competent
technicians must have other attributes.
They must be determined to do a safe job
under any circumstances. They must be
willing to listen and to cooperate with the
many types of personnel encountered in
the field, but they must not compromise
the safety aspects of their work for the
convenience of themselves, their crew or
someone else.
Need for Safety Training of Test
Personnel
Test personnel can acquire a proper
attitude and point of view toward safety
only through training coupled with
experience. The training program should
include first aid and lifesaving techniques.
In situations where irritating, toxic or
corrosive dusts, gases, vapors or fluids are
present, test technicians should be given
special training to make sure that they are
familiar with the properties of these
substances and with the methods of
controlling the hazards. Emergency
procedures must be learned and test
personnel must know where medical and
hospital assistance is available at all hours.
Leak testing technicians should have
more thorough training in accident
prevention than the regular plant or
construction workers. For leak testing
personnel, safety involves not a set
pattern of activity but a complex and
constantly changing set of problems.
102
Leak Testing
The United States Department of
Transportation is responsible for the rules
governing training requirements for
handlers of hazardous materials
(HAZMAT). The Code of Federal
Regulations1 states the requirement that
hazardous materials handlers receive
training at least every two years by
someone licensed to provide such
training.
Hazards in Leak Testing
Precleaning of test surfaces is required for
leak testing where surface contamination
might prevent entry of fluid tracers. Many
cleaning processes involve liquid solvents
and vapors, some of which present
possible hazards of flammability, toxicity
or asphyxiation. Liquid leak tracers often
have similar hazards, if vapors accumulate
in working areas. Ventilation must be
provided to prevent hazardous vapor
concentrations. Electrical systems must be
properly grounded and enclosed or
protected to prevent ignition of
flammable vapors in air. Access to test
surfaces, particularly on large structures,
can be hazardous if scaffolding is
inadequate, lighting is insufficient or bad
housekeeping creates hazards such as oily
work surfaces or obstructions in
passageways.
Special Safety
Considerations in Testing
Systems under Pressure
When a pressure or a vacuum vessel is
fabricated, some means of testing this
vessel must be used to predict safe
performance. It is sometimes necessary to
exceed the designed operating conditions
during initial pressure testing. This
pressurization requires many safety
considerations to ensure proper protection
of personnel. Greater respect for high
pressure has led to increased safety
emphasis, with the result that overall
safety experience has been good.
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
Psychological Factors and
the Safety Program
disasters. In today’s industry it is the
responsibility of the employer to provide
employees adequate training on safety
practices for for their job responsibilities.
The nature of leak testing work dictates
that a competent safety program be used.
Much of the success of such a program
depends on its acceptance by those to
whom it is directed. Never has there been
a safety device or a safety program that
some human being could not disrupt or
impair. The human factors that operate at
all levels in industry are perhaps the most
potent factors for success or failure of a
safety program. The president of a
company, the safety director and the leak
testing supervisor may either emphasize
safety or subordinate it to production
goals. Production, maintenance and
testing personnel are also important
contributors to safety and their full
cooperation is vital. Individual differences
affect personnel acceptance of a safety
program. These differences must be
recognized when motivating work groups
to use good safety practices at all times.
The safety program must be designed
with an understanding of motivation of
people. To want something is to be
motivated, but not to want something
also requires motivation. To use a safety
device to protect one’s fingers from a saw
shows motivation for safe practice.
However, the desire to ignore a safety
device that interferes with production is
caused by still other motives. Conflicting
motivations should also be considered in
any attempt to understand human
relations that influence the success of
safety programs. Industry has recognized
the effects that attitudes can have on
production, plant morale and plant safety.
As a result, management should spend
considerable effort to determine the
attitudes of its workers. Measuring,
developing and changing attitudes
constitute a major problem for personnel
and psychologists and are of extreme
importance to the safety program.
Personnel Safety Training
Requirements
There should always be concern with
safety training of personnel. The learning
process starts at birth. Most early safety
training is through experience, as when a
child may have touched a stove and been
burnt, played with a knife and been cut or
fallen from a precarious treehouse and
broken a bone. However, personnel
testing today’s vessels that hold gases,
vapors and liquids at various temperatures
and at pressures ranging from high
vacuums (in nanopascal) to high pressures
(in megapascal) cannot afford to learn
safety by causing or experiencing
Safety Aspects of Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
103
PART 2. Control of Hazards from Airborne Toxic
Liquids, Vapors and Particles
Toxic Gas and Vapor
Sensors and Alarms
Detection and warning of the presence of
toxic vapors or gases in a work area can be
provided by various types of electronic
instruments with detectors and alarm
systems responsive to many different
airborne chemicals, fumes, smoke or
particulate matter. For general protective
service applications, wall mounted,
self-contained monitors can detect and
provide audible signals of the presence of
various combustible gases, fumes and
microscopically sized airborne particulate
contaminants. These are typically
provided with pilot lights to indicate the
presence of alternating current line power
and standby battery power. Flashing red
lights actuated when abnormal
concentrations of contaminants occur.
The alarm sensitivity control can be
adjusted to allow compensation for the
normal ambient quiescent atmospheric
contamination levels.
The sensor assembly of a typical gas
monitor and alarm system contains a
heated semiconductor element whose
resistance to current flow varies as a
function of the type and quantity of gas
molecules adsorbed on its surface. The
heater effectively boils off adsorbed
contaminants. The sensor resistance is
thus primarily a function of the adsorbed
gas molecules, whose number is related to
their relative concentrations in the
ambient air atmosphere. The sensor is
designed for more than 50 000 exposures
and can detect 50 µL·L–1 of many
combustible and toxic gases and vapors,
including those listed in Table 1.
Selecting Leak Testing Sites with
Adequate Ventilation
When possible, testing of structures such
as pressure vessels should be performed in
a well ventilated area isolated far from
other processes such as welding or
grinding. A room is desirable with a high
roof, adequately ventilated at its apex and
with enough low level inlets. Conversely,
a small room with a low roof and a
minimum of opening for ventilation
should not be used for testing with
potentially dangerous tracer gases such as
hydrogen.
104
Leak Testing
Ventilation to Reduce
Vapor Hazards in Solvent
Use Areas
Many applications of leak testing in
various industries have, as a prerequisite
to testing, some cleaning operation. This
operation often uses volatile solvents that
can contaminate the air within
enclosures; therefore, some consideration
must be given to ventilating the working
areas with explosion-proof equipment.
Local exhaust systems have several
inherent advantages over general
ventilation for removal of atmospheric
contaminants. They permit removal of
hazardous vapors before they spread
throughout the work area, they provide
economy of air flow and they involve less
heat loss. Local exhaust systems are
impractical where the contaminant is
usually a solvent vapor. Local exhausts
may be unsuitable because there are a
multitude of sources of vapor, or the
source may be extensive, or the amount
of ductwork to connect all the necessary
hoods may be too costly or impractical.
The basic purpose of volatile solvents
used in industrial cleaning operations is
to dissolve or loosen contamination such
as grease, dirt and other impurities and so
facilitate their removal. The solvent may
tend to evaporate into the atmosphere.
This evaporation of volatile constituents
leaves behind some physically changed
substance that must be removed from test
surfaces. Thus, the use of solvents in these
processes involves polluting the air with
vapor. The aim of the safety engineer is to
keep this vapor concentration as low as
possible, certainly below the toxic limit. If
local exhaust systems are inadequate, such
widely distributed solvent vapors can
sometimes be controlled by diluting the
general room atmosphere with outdoor
air fast enough to keep the concentration
of toxic vapor in the air of the working
space within safe limits.
Ventilation Rate
Calculations for Safe Use
of Vaporizing Solvents
The rate of solvent evaporation can easily
be ascertained, as can the chemical nature
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
of the solvent. It is known that the weight
of a given volume of vapor that
evaporates from a liquid is proportional to
its molecular weight. It is possible, then,
to calculate how much air must be mixed
with a solvent vapor to hold the
concentration down to safe limits.
Table 2, from which general ventilation
can be calculated, is based on the formula
of Eq. 1 (in SI units):
(1)
VR
=
(2.4
× 10 7
) WM (VDC)
where VR is rate of ventilation (m3·min–1);
W is rate of solvent evaporation
(kgm·min–1); M is molecular weight of
solvent (unified atomic mass unit); and
VDC is ventilation design concentration
(from Table 2).
Equation 1 does not give the
maximum acceptable concentration for
the compound. Instead, it is the
ventilation design concentration that has
incorporated in it a safety factor based on
toxicity, order and experience. Equation 1
converts to Eq. 2 in English units:
(2)
VR
=
(4
× 10 8
) WM (VDC)
where VR is rate of ventilation (ft3·min–1);
W is rate of solvent evaporation
(lbm·min–1); M is molecular weight of
solvent (unified atomic mass unit); and
VDC is ventilation design concentration
(from Table 2).2
Neither the maximum allowable
concentration (MAC) nor the threshold
limit value (TLV) should be used for
calculating the ventilation design
concentration. The degree of vapor
dilution in the working space is bound to
be uneven. In addition, the
concentrations must always be
maintained below the MAC or TLV to
provide a factor of safety. In turn, this
factor of safety depends on whether the
solvent vapor is to be controlled because
TABLE 1. Combustible and toxic gases and vapors detectable by area monitors and
alarm systems.
Acetaldehyde
Acetone
Acetonitrile
Acetylene tetrabromide
Alcohol
Allyl alcohol
c-allylglycidylether
Ammonia
Benzene
Benzoyl chloride
Benzoyl peroxide
Butane
2-butanone (MEK)
2-butoxyethanol
Butyl acetate
Butyl alcohol
Camphor
Carbon monoxide
Carbon tetrachloride
Chloroacetaldehyde
Chlorobenzene
c-chloroform
1-chloro-1-nitropropane
Chloropicrin
Chloroprene
Cumene
Cyclohexane
Cyclohexanol
Cyclopentadiene
DDT
Diacetone alcohol
Diazomethane
Diborane
1,1 dichloroethane
1,2 dichloroethane
Diethylamine
Diethylamino ethanol
Diisobutyl ketone
Dimethylamine
Dimethylaniline
Dimethylformamide
1,1 dimethylhydrazine
Dinitrobenzene
Dinitrotoluene
Dipropylene glycol
methyl ether
Epichlorhydrin
2-ethoxyethanol
Ethyl alcohol
Ethylamine
Ethyl benzene
Ethyl bromide
Ethyl butyl ketone
Ethyl chloride
Ethyl ether
Ethyl formate
Ethylenediamine
Ethyl dichloride
Ethylene oxide
Formaldehyde
Furfuryl alcohol
Gasoline
Glycol monoethyl ether
Heptane
Hexachloroethane
Hexane
2-hexanone
Hexone
Hydrogen
Hydrogen bromide
c-hydrogen chloride
Hydrogen cyanide
c-hydrogen sulfide
Isoamyl alcohol
Isobutyl alcohol
Isopropyl alcohol
Ketone
Liquid propane gas
Methane
Methyl acetylene
Methylal
Methyl alcohol
Methylamine
Methyl n-amyl ketone
Methyl butyl ketone
Methyl cellosolve
Methyl chloride
Methyl chloroform
Methylcyclohexane
Methylcyclohexanol
Methylene chloride
Methyl ethyl ketone
c-methyl mercaptan
Naphtha
Naphthalene
Natural gas
Nitrobenzene
p-nitrochlorobenzene
Nitroethane
Nitroglycerin
Nitromethane
Nitrotoluene
Ozone
Pentane
2-pentanone
Perchloroethylene
Petroleum distillate
Phenylether
Propane
Propargyl alcohol
Propylene oxide
Propyne
Refrigerant-11, -134a etc.
Steam
Stibine
Sulfur dioxide
Sulfur hexafluoride
Tetrachloronaphthalene
Tetranitromethane
Toluene
1,1,1 trichloroethane
1,1,2 trichloroethane
Trichloroethylene
Trichloronaphthalene
1,2,3 trichloropropane
Trinitrotoluene
Turpentine
Xylene
Safety Aspects of Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
105
of its inherent toxicity or its disagreeable
odor.
Example of Ventilation Rate
Calculation
For example, suppose that 3 L of methyl
ethyl ketone were evaporated per hour.
One liter of methyl ethyl ketone requires
a ventilation quantity of 1800 m3 of air;
3 L would then require 3 times 1800
equals 5400 m3 of air. If this is needed per
hour, the ventilation rate per minute
would be 5400 divided by 60 equals
90 m3·min–1. (Users of the English system
should use a conversion of 35 ft3·m–3 and
2 pt·L–1, or 17 ft3·pt–1 for each m3·L–1.)
It is important to note that this
example assumes there is perfect mixing
of the clean air with the solvent vapor,
but in practice this does not occur. The
ventilation rate calculated is therefore a
minimum. It should be increased
depending on other factors involved, such
as type and location of air diffusers,
location of people in the working space
and relative toxicity of the vapor.
The volume of the space in which the
work is done does not enter the
calculation for ventilation design
concentration. This is a variance from the
common practice of specifying ventilation
requirements in terms of number of air
changes per minute, which of course
directly involves the work space volume.
The rule of thumb based on room air
changes per minute, thus in widespread
use over many years, has been used
improperly more often than properly. This
is especially true when there are
unwanted contaminants being released
within the space.
Example of Evaluation of Health
Hazard from Dilution Rate Table
The following is an example in which the
degree of health hazard resulting from a
solvent exposure is to be evaluated using
data from Table 2. Trichloroethylene is
being used in an enclosed 6 × 6 × 3 m
work space. In an 8 h day, 20 L of the
solvent are lost through evaporation.
There are two air changes per hour. Is
there a potential health hazard?
Solution in metric units. The work
space volume is 6 × 6 × 3 = 108 m3.
Ventilation rate at two changes per hour
provides 2 × 108 = 216 m3·h–1. The rate of
solvent evaporation is 20/8 = 2.5 L·h–1.
The dilution rate or ventilation ratio is
216 divided by 2.5 = 86 m3·L–1. The
proper ventilation ratio (from Table 2)
should be 2700 m3·L–1. Therefore, the
ventilation rate is totally inadequate and a
health hazard is indicated. At least 2700
divided by 86 is 31 times as much
ventilation is required for the safe
TABLE 2. Dilution rates for common industrial solvents recommended for use in ventilation design (SI units), after
Hemeon.2
Solvent
Acetone
Benzene
Carbon tetrachloride
Ether
Ethyl alcohol
Isopropyl alcohol
Methanol
Methyl-ethyl ketone
Pentachloroethane
PMV naphtha
Stoddard solvent
Tetrachloroethane
Tetrachloroethylene
Toluene (toluol)
Trichloroethane
Trichloroethylene
Xylene (xylol)
Molecular
Weighta Densityb VDCc
(M)
(kg·m–3) (µL·L–1)
58
78
154
74
46
60
32
72
202
110
130
168
166
92
133
131
106
790
880
1580
720
790
790
800
810
1670
750
800
1580
1620
870
1440
1460
880
150
——e
——e
75
250
150
100
150
——e
200
500
5
100
100
100
100
75
Ventilation Ratio or Dilution Rated
(Quantity of Air per Unit, Solvent)
(m3·kg–1) (m3·L–1)
2800
——e
——e
4300
2100
2700
7500
2200
——e
1100
370
29 900
1400
2600
1800
1800
3000
2200
——e
——e
3000
1600
2100
6000
1800
——e
300
300
45 000
2300
2300
2600
2700
2700
(ft3·lb–1) (ft3·pt–1)
46 000
——e
——e
72 000
34 000
45 000
125 000
37 000
——e
18 000
6000
480 000
24 000
44 000
30 000
30 000
50 000
38 000
——e
——e
54 000
28 000
37 000
103 000
37 000
——e
14 000
5000
790 000
40 000
39 000
45 000
45 000
46 000
Possible
Complaints
If Twice VDCc
Exceeded
Disagreeable
——e
——e
Disagreeable
Disagreeable
Disagreeable
Toxic
Disagreeable
——e
Disagreeable,
Disagreeable
Toxic
Disagreeable,
Toxic
Disagreeable,
Disagreeable,
Disagreeable
toxic
toxic
toxic
toxic
a. Atomic mass units.
b. Same as g·L–1 or mg·cm–3.
c. Ventilation design concentration, not to be identified with values of maximum acceptable concentration or threshold values employed in appraising
conditions since all VDCs include a factor of safety.
d. Ventilation ratio (or dilution rate) is the ratio of the volume of air (m3 or ft3) to the volume or weight of solvent evaporated.
e. Dilution system is not recommended in this case.
106
Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
operation of this facility. Note that this
type of calculation is valid only if the air
contaminant is uniformly distributed at a
relatively low concentration. Where the
air contaminant is localized in high
concentrations, more complex means of
evaluating the hazard must be used.
For users of the English system, the
preceding example could be stated as
follows. Trichloroethylene is used in a
room of 20 × 20 × 10 ft. In an 8 h day,
5 gal are evaporated; there are two air
changes per hour. The solution in English
units is, for room volume, 20 × 20 × 10 =
4000 ft3; for ventilation rate, 2 × 4000 =
8000 ft3·h–1.
Rate of solvent evaporation = 5 divided
by 8 equals, in United States units,
0.6 gal·h–1 or 5 pt·h–1.
Ventilation ratio is 8000 divided by 5 is
1600 ft3·pt–1. Ventilation ratio according
to Table 2 is 2700 m3·L–1 or 17 times 2700
is 46 000 ft3·pt–1. At least 46 000 divided
by 1600 = 29 times more ventilation is
required.
Evaluation of Toxicology
and Health Hazards of
Materials
The toxicity of a material is not
synonymous with its health hazard.
Toxicity is the capacity of a material to
produce injury or harm. Hazard is the
possibility that a material will cause injury
when a specific quantity is used under
specific conditions. The key elements to
be considered in evaluating a health
hazard are the following.
1. How much of the material is needed
to produce injury?
2. What is the probability that the
material will be absorbed by the body
to produce injury?
3. What protective equipment is in use?
Because toxicity is not a definite
physical constant but rather the degree to
which a substance will affect living cells
under certain conditions, it can be
measured only after recognizable changes
have occurred following absorption. Some
changes such as impaired judgment or
delayed reaction time may be produced at
levels too low to cause actual cell damage.
Then too, toxicity depends on the dose,
rate, means and site of absorption. Other
pertinent factors include the ambient
temperature and the working conditions,
as well as the general state of health,
individual differences, tolerance and diet
of individual personnel.
Estimating Toxicity Values
and Lethal Doses of Toxic
Materials
The first attempts at estimating the
toxicity of a substance are usually made
on the basis of animal experiments. Data
from these experiments are expressed as
lethal doses (LD) in milligrams of
substance per kilogram of body weight of
the test animal. The commonly used
expressions are the following: MLD,
minimum lethal dose, the smallest dose
that kills one of a group of test animals;
LD50, lethal dose for 50 percent, the dose
that kills one half of a group of test
animals (usually ten or more); LD100,
lethal dose for 100 percent, the dose that
kills all of a group of test animals (usually
ten or more). These doses may also be
expressed as lethal concentrations (LC) for
airborne toxic substances.
Substances can then be rated according
to their relative toxicity as shown in
animal experiments (Tables 3 and 4).4-6
The probable lethal dose for humans is
often estimated from animal tests. These
ratings are based on the results of short
term exposures only. It is possible in
TABLE 3. Combined tabulation of toxicity classes, after Roehrs and Center.3
Commonly
Used Term
LD50 Single
Oral Dose for Ratsa
(g·kg–1)
Extremely toxic
Highly toxic
Moderately toxic
Slightly toxic
Relatively nontoxic
Practically nontoxic
≤0.001
0.001 to 0.05
0.05 to 0.5
0.5 to 5.0
5.0 to 15.0
>15.0
4 h Vapor Exposure
Causing 2 to 4 Deaths in
Six-Rat Group
(µL·L–1)
>10
10
100
1000
10 000
>100 000
to
to
to
to
100
1000
10 000
100 000
LD50 Skin Exposure
for Rabbits
(g·kg–1)
≤0.005
0.005 to 0.043
0.044 to 0.340
0.35 to 2.81
2.82 to 22.6
>22.6
Probable Lethal
Dose for Humans
_________________________
SI
(English)
50 mg (taste) (1
4 cm3
(1
30 cm3
(1
0.5 L
(1
1L
(1
>1 L
(>1
grain [taste])
tsp)
oz)
pt)
qt)
qt)
a. Grams of dose per kilogram of rat.
b. Parts of vapor in million parts of air.
Safety Aspects of Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
107
actual, long term chronic exposure for a
substance to prove highly toxic, even
though short term exposure tests
indicated a low order of toxicity. However,
animal experiment data are difficult to
interpret and apply to human exposures.
Such data are valuable only as guides to
be used in estimating the gross toxicity of
a substance and as leads for further
investigations.
NIOSH Evaluations of Exposure to
Toxic Substances
In the United States, the Department of
Health, Education and Welfare (HEW), the
Occupational Safety and Health
Administration (OSHA) and the National
Institute for Occupational Safety and
Health (NIOSH) conduct critical reviews
of occupational hazards, prepare criteria
documents, recommend standards of
exposure and list toxic effects of chemical
materials. Under no circumstances can the
toxic dose values presented for chemical
substances be considered as being
definitive values for describing safe versus
toxic doses for human exposure.
Concentrations of chemical substances in
the work environment that may be safely
tolerated can be determined only by a
critical evaluation of all available
pertinent data by experienced
investigators.
NIOSH special occupational hazard
reviews analyze and document, from a
health standpoint, the problems
associated with a given industrial
chemical, process or physical agent and
recommend the implementation of
engineering controls and work practices
to relieve these problems. The evaluations
pertain primarily to special alleged
hazards, e.g., those with carcinogenic,
mutogenic, teratogenic or other
reproductive effects, although they may
review other effects as needed.
The permissible exposure levels of
hazardous substances that have been
adopted by OSHA to provide a safe,
healthful work environment for all
persons are cited as Occupational
Standards (OSHA). These are given in an
annually updated NIOSH Registry of Toxic
Effects of Chemical Substances. NIOSH
Criteria Documents contain
environmental and medical
recommendations related to specific
substances and processes.
Management and test personnel can
use NIOSH published resources to
determine probabilities of hazards with
new test materials, interacting
combinations of chemical materials and
environmental hazards. In all cases of
doubt, however, reference to experts in
the field for consultation and guidance is
recommended.
Limitations of Safety
Warnings
This volume is limited to leak testing and
endeavors to provide comprehensive and
useful information and data on test
techniques and applications. It is not
possible, within its scope, to advise users
of all potential hazards and toxic or
dangerous substances. In this book, only
partial information and warnings can be
included, so workers and test personnel or
management should look up more
complete data in publications from
NIOSH and other sources for complete
information. Qualified assistance should
TABLE 4. Guidelines for evaluating acutea dosages differentiating relatively toxic from nontoxic substances taking into
consideration the route of administration to experimental animals and the dose causing deathb. After Hine and
Jacobson5 and NIOSH 78-104A.6
Species
Rectal
24 h
Subcutaneous
Intraduodenum Inhalation
Intraperitoneal Intradermal
Intracervix
Maximum Skin Intrapleural
Implant
(mg·kg–1)
(g·kg–1) (g·kg–1) (g·kg–1)
(g·kg–1)
Frog, gerbil, hamster
Mouse, rat, squirrel
Bird, chicken, duck, guinea,
pig, pigeon, quail, rabbit, turkey
Cat, cattle, dog, goat, horse,
monkey, pig, sheep
Other
Unspecified
Parenteralc Parenteral Unreported
(g·kg–1)
(g·kg–1)
(g·kg–1)
2.5
5.0c
10.0
1.0
2.0
4.0
1.4
2.8
2.8c
1.0
2.0
4.0
5.0
10.0c
20.0
0.75
1.5
3.0
1.0
2.0
4.0
2.5
5.0
10.0
10.0
4.0
5.6
4.0
20.0
3.0
4.0
10.0
a. Applies to those substances for which acute or short-term toxicity characterizes the response, e.g, fast-acting substances, irritants, narcosis-producing
substances, and most drugs. Does not apply to substances whose characteristic response results from prolonged exposures, e.g., silica, lead, benzene,
carbon disulfide, carcinogens. Concentrations more appropriately characterizing the toxicity of long- or slow-acting substances are derived from nonacute
toxicity studies.
b. Calculated from experimental data (Stokinger).
c. Intravenous, intramuscular, ocular, intracerebral, intratracheal, intraplacental, intravaginal, intrarenal.
108
Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
be sought from experts in safety, legal
requirements, governmental regulations,
safety engineering, health and medical
practice, wherever the possibility of
hazards may exist. Special reference
should be made by leak testing personnel
and supervision to applicable plant safety
rules; to procedures used in case of
accidents; to local, municipal, county,
state and national laws and regulations;
and to qualified safety and health
agencies, organizations and experts for
advice on health and safety. The warnings
and precautions given in this book are
based on experience in industry during
application of leak tests. They do not
foresee the possibilities and nature of
potential future accidents, nor do they
include the constantly changing
identifications of toxic or hazardous
substances included in publications of
governmental and other health and safety
agencies and organizations.
Precautions with Specific
Fluids
Acetone and Other Ketones
Acetone and other ketones are typical
solvents and metal cleaning compounds
used widely in industry. Acetone
(dimethyl ketone) is a very flammable
liquid that should be handled and stored
with precautions against fire and
explosion. In spite of the large quantities
of acetone used in industry and its high
volatility, there are no known
documented reports of serious industrial
poisoning. Experimental work has shown
that acetone is a narcotic. Overexposure
will lead to moderate irritation of the
eyes, nose and throat and to headache,
stupor and a general feeling of oppression.
The absorbed acetone is eliminated slowly
and the symptoms are persistent. Contact
with skin and eyes should be avoided by
the use of protective clothing. In areas of
vapor concentration, approved respiratory
protective equipment should be used.
Precautions with Halogenated
Hydrocarbons
Halogenated hydrocarbons are typically
colorless volatile liquids with excellent
organic solvent properties and are widely
used. Hydrocarbons having only one or
two halogens are usually flammable and
less toxic than similar hydrocarbons with
complete halogen substitution. Thermal
decomposition of halogenated
hydrocarbon vapors occurs and poisonous
gases may be formed when they come
into contact with a heat source, such as a
red hot surface, flame or electric arc.
The most common halogenated
hydrocarbons, arranged in increasing
order of ability to produce narcosis, are
vinyl chloride, methyl chloride, ethyl
chloride, ethylene dichloride, ethyl
bromide, carbon tetrachloride,
dichloromethane (also called methylene
chloride), methyl chloroform (also called
1,1-trichloroethane and
1,1,2-trichloroethane), trichloroethylene,
methyl bromide, tetrachloroethylene (also
called perchloroethylene),
pentachloroethane and tetrachloroethane
(see a dictionary of commercial
chemicals). Tetrachloroethane is about 40
times as strong a narcotic as vinyl
chloride. An acute exposure to the more
narcotic of these compounds may result
in unconsciousness for a surprisingly long
period, with eventual recovery.
Unconsciousness for eight weeks has been
reported in a case of methyl bromide
poisoning. It is to be noted that the
preceding listing is not in the same order
as the chronic toxicity of these
halogenated hydrocarbons. Chronic
toxicity due to low rates of exposure over
long periods of time has been the more
common problem in industry.
Tetrachloroethane, the most toxic of
the common chlorinated hydrocarbons,
has no particular warning signs or
symptoms. It can produce extremely
severe poisoning from continuous
exposure to fairly low concentrations.
Tetrachloroethane is a very dangerous
compound because inhalation of it at a
concentration barely perceptible by odor
can lead to extensive injury. Carbon
tetrachloride, methyl chloride,
dichloroethylene and trichloroethylene
show decreasing chronic toxicity in
approximately that order. Introduction of
a bromine or iodine atom into one of the
halogenated hydrocarbons generally
increases the toxicity as compared to that
of the corresponding chlorine compound.
In contrast, introduction of a fluorine
atom generally reduces the toxicity as
compared to that of the corresponding
chlorine compound.
The methyl compounds, particularly
methyl chloride and methyl bromide, are
in a special class because of their delayed
action. Minor symptoms may appear
during an acute exposure to these
compounds; severe symptoms may appear
after a delay of several hours to several
days.
Precautions with Carbon
Tetrachloride
Carbon tetrachloride (CCl4) is a
halogenated hydrocarbon liquid that is
colorless, nonflammable and has a
characteristic odor. Synonyms for carbon
tetrachloride include tetrachloromethane
Safety Aspects of Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
109
and perchloromethane. Carbon
tetrachloride is used as a solvent,
degreaser and chemical constituent and
can act to remove the natural liquid cover
of human skin. With repeated contact
with the skin, it can lead to a dry, scaly,
fissured skin condition known as
dermatitis. Chronic poisoning including
liver damage comes from long, continued
absorption of fairly small amounts of
carbon tetrachloride over a long period.
Barrier creams, gloves, protective clothing
and masks should be used as appropriate
where exposure occurs. The major
problem in prevention of injuries from
carbon tetrachloride is that of prevention
of inhalation of carbon tetrachloride
solvent vapor. Oxidative decomposition
by flame causes it to form phosgene (a
poisonous gas) and hydrogen chloride,
also a poisonous gas. Carbon tetrachloride
is now prohibited in many instances.
Precautions with Fluorocarbon
and Refrigerant Gases
Fluorocarbons are hydrocarbons
containing fluorine; they may contain
other halogens in addition to fluorine.
Generally these compounds are colorless
nonflammable gases. Decomposition of
chlorine-containing fluoromethanes,
caused by contact with an open flame or
hot metal, produces hydrogen chloride,
hydrogen fluoride, phosgene, carbon
dioxide and chlorine.
The fluorocarbons are used primarily as
refrigerants, leak testing tracer gases and
fire extinguishers and in degreasing of
electronic equipment. They have found
wide use due to their relatively low
toxicity and nonflammability. Trademarks
including Freon®, Genetron® and Isotron®
have been used for a number of
fluorocarbons used in refrigeration. The
fluorocarbon compounds may produce
mild irritation in the upper respiratory
tract, perhaps caused by their
decomposition products. Dermatitis
occurs only rarely from contact with these
materials.
In the United States, the
Environmental Protection Agency took
action to essentially ban the
chlorofluorocarbons in aerosol spray cans
that release the chemicals to the
atmosphere with each use of the can. The
law itself is written in two parts, which
are integrated. The first part is
administered by the Food and Drug
Administration. The second part, which
covers penetrants, is administered under
the Toxic Substances Control Act. The
exact wording appears in Parts 712 and
762 of this act and in the Federal Register
of March 17, 1978. The important
wording appears in paragraph 762.12(a),
as follows: “After December 15, 1978 no
110
Leak Testing
person may process any fully halogenated
chlorofluoroalkane into any aerosol
propelled article. . . .”
Prevention of Personnel Exposures
to Halogenated Hydrocarbons
Because exposure of testing personnel to
halogenated hydrocarbons is almost
invariably by inhalation, the most
valuable measures to prevent poisoning
are enclosure and ventilation at the point
where vapor is released. However, several
of the chlorinated hydrocarbons are
apparently much more toxic by skin
contact than has been believed. Skin
contact should therefore be avoided
because of the probability that where
there is skin contact there will also be a
severe inhalation exposure.
Precautions with Aromatic
Hydrocarbons
Aromatic hydrocarbons are widely used as
solvents and chemical intermediates. The
basic aromatic nucleus is benzene, C6H6.
Because of its health hazards, benzene has
been replaced as a commercial solvent by
toluene and other less toxic compounds.
Typically, the vapor of aromatic
hydrocarbons causes central nervous
system depression and other effects. Vapor
is absorbed through the lungs and the
liquid may be absorbed through the skin.
Repeated and prolonged skin contact may
cause defatting of the skin, which leads to
dermatitis. Chronic benzene poisoning
can be fatal.
Precautions with Methyl Alcohol
Methyl alcohol (CH3OH) is a colorless,
volatile liquid with a mild odor. It is used
in synthesis of many chemicals and as an
industrial solvent. Contact of methyl
alcohol with the skin can produce mild
defatting and a mild dermatitis that can
be avoided by use of barrier creams and
protective clothing. Methyl alcohol is
virtually nonirritating to the eyes or upper
respiratory tract at concentrations in air
below 2000 µL·L–1; it is difficult to detect
by odor at less than this level.
Methanol (methyl alcohol) poisoning
is usually produced by swallowing the
liquid or inhaling high concentrations of
vapor in an enclosed area. The signs of
poisoning include headache, nausea,
vomiting, violent abdominal pains,
aimless and erratic movements, dilated
pupils, sometimes delirium and such eye
symptoms as pain, tenderness on pressure
and, occasionally, blindness. Direct action
of the liquid or the vapor on the skin and
mucous membranes may produce an
irritation and inflammation.
One of the peculiarities of methanol
poisoning is its exceptionally severe action
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
on the optic nerve. About one half of all
the serious cases of methanol poisoning
result in some impairment of vision. This
loss is usually permanent and may vary
from dimness or blind spots scattered
through the visual field to total blindness.
Precautions with Glycols and
Glycol Derivatives
Glycols are dihydric alcohols, which are
colorless, odorless liquids. Glycols are
soluble in water and in alcohol, have high
boiling points, have low freezing points
and are used as solvents and antifreeze.
These compounds have relatively low
toxicity and the major hazard appears
when the liquids are heated during
processing.
Precautions with Ethylene Glycol
Ethers
Ethylene glycol ethers are only mildly
irritating to the skin. Vapors may cause
conjunctivitis and irritation of the upper
respiratory tract. Temporary corneal
clouding may also result and may last
several hours. Acetate derivatives cause
greater eye irritation than the parent
compounds. The butyl and methyl ethers
may penetrate the skin readily. Symptoms
from repeated overexposure to glycol
ether vapors are fatigue and lethargy,
headache and tremor. Glasses and
protective clothing can be used to prevent
skin absorption. Respiratory protection
maybe needed if ventilation is poor or
glycol compounds are heated or atomized.
Precautions with Petroleum
Derivatives
Naphtha is a rather indefinite term for
any one of a number of solvent mixtures
derived from petroleum. One should
define it more carefully before attempting
to assess the hazard. The naphthas are
irritating to the skin, conjunctiva and
mucous membranes of the upper
respiratory tract. Skin chapping and
photosensitivity may develop after
repeated contact with liquid naphtha. If
confined by clothing against the skin, the
naphthas may cause skin burn. Workers
should use barrier creams, protective
clothing, gloves and masks where
exposure to naphtha vapor is likely.
Sufficient quantities of naphtha cause
central nervous system depression.
Symptoms include inebriation, followed
by headache and nausea. In severe cases,
dizziness, convulsions and
unconsciousness may result. If benzene is
present, coal tar naphthas may produce
leukemia.
Precautions with Stoddard Solvent
Stoddard solvent is a registered
commercial standard of the U.S.
Department of Commerce for a dry
cleaning solvent. Its specifications are that
it has a flash point of 37.8 to 43.3 °C
(100 to 110 °F), evaporates without
residue and consists of aliphatic, saturated
materials and, in some formulations, 15
to 20 percent aromatics. The fire hazard is
about that of kerosene. It is available
under a number of trade names.
Precautions with Toluene
Toluene is seldom a source of acute
poisoning, although its inherent acute
toxicity is somewhat higher than that of
benzene. It is a flammable, colorless liquid
of rather strong aromatic odor that serves
somewhat as a warning of high
concentration. At concentrations of 500
to 1000 µL·L–1, toluene is strongly
irritating to the eyes and respiratory
system. In higher concentration, it is a
narcotic and the signs of acute poisoning
are headaches, drunkenness, nausea,
vomiting and ultimately unconsciousness.
Toluene does not appear to produce the
severe and often fatal depression of the
blood forming organs seen in chronic
benzene poisoning. In case of acute
exposure to toluene, the person should be
taken to fresh air as soon as possible.
Oxygen should be given and, if breathing
has stopped, artificial respiration should
be administered immediately. A physician
should be called at once.
Precautions with
Trichloroethylene
Trichloroethylene is a halogenated
hydrocarbon used primarily as a
degreasing compound. It has no flash
point as such, but at elevated
temperatures and with a high energy
ignition source, such as a welding arc, its
vapors can and will explode. Toxic
decomposition products, mainly
hydrogen chloride with some phosgene,
both highly poisonous gases, may also be
formed under these conditions. Phosgene
may be formed inside a cigarette when
smoking in an area where
trichloroethylene vapors are present.
Trichloroethylene may have a
depressant action or, as with other
chlorinated hydrocarbons, cause
alteration of the heart rhythm, or lead to
addiction. Although some absorption may
occur through the skin, trichloroethylene
has mainly a defatting and dermatitisproducing skin effect.
Safety Aspects of Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
111
Precautions with Xylene
Xylene, C6H4(CH3)2, is a mixture of
isomers and may contain numerous other
solvent compounds. It is used as a solvent
and is specified in some tests to detect the
water content of penetrant materials.
Xylene vapor may cause irritation to the
eyes, nose and throat. Repeated or
prolonged skin contact may cause drying
and defatting of the skin, which may lead
to dermatitis. Liquid xylene is irritating to
the eyes and mucous membrane.
Aspiration of a few milliliters may cause
severe effects. Repeated exposure of the
eyes to high concentrations of xylene
vapor may cause irreversible eye damage.
When xylene vapor concentrations
exceed allowable standards, full face
masks with organic vapor cannisters or air
supplied respirators should be furnished.
Impervious protective clothing and gloves
should be worn by personnel exposed to
liquid xylene. Xylene wet clothing should
be changed quickly. Goggles or safety
glasses are advised. Barrier creams may be
useful.
Hazards of Oxygen Deficient
Atmospheres
Oxygen deficiency designates an
atmosphere having less than the
percentage of oxygen found in normal air.
Normal air contains about 21 percent
oxygen at atmospheric pressure. When
the oxygen concentration in air is reduced
to approximately 16 percent, many
individuals become dizzy, experience a
buzzing in the ears and have a rapid
heartbeat.
In addition to tests for toxicity, the
oxygen content of the atmosphere of a
vessel or similarly confined space
suspected of being oxygen deficient
should be determined by preentry and
subsequent tests made with instruments
approved for the purpose by the United
States Bureau of Mines. No one should
enter or remain in a vessel or enclosed
space that tests show has less than
16 percent oxygen in its atmosphere at
any time unless wearing approved
respiratory protective equipment such as a
fresh air hose mask or self-contained or
self-generating breathing apparatus.
Various types of self-contained
compressed air breathing apparatus,
approved by the U.S. Bureau of Mines,
have proved satisfactory in oxygen
deficient atmospheres. They are especially
useful where it is difficult to run an air
supply hose line.
two toxological effects from this inert gas
are asphyxiation and radiation exposure.
To satisfy federal and state licensing
requirements in the United States, the
pressurization systems are provided with a
room enclosure; an exhaust system for
typically 3 to 5 min room air exchange; a
series of interlocking safety circuits for the
proper exhaust air flow; and the detection
of any radioactive gas in the room or
exhaust.
The regulatory agencies monitor and
enforce these requirements as well as
continuous monitoring film badges to
document worker and room exposure
levels. A total dump of a typical
krypton-85 leak testing system would
require immediate operator evacuation of
the machine enclosure, which would
typically result in nondetectable radiation
exposure as measured on state-of-the-art
film badges.
Precautions with Dry
Powder Developers
Dry powder developers as used in some
liquid leak tracers are subject to dusting
and other behavior characteristics of dry
powder materials. Safety procedures such
as the following should be observed.
1. Avoid continued excessive inhalation.
2. Use a well fitting dust mask and
adequate ventilation.
3. Wear eye protection when filling or
emptying a hopper.
4. Any dry powder material can build
static electricity charges when
subjected to the friction of mixing,
sliding or conveying. Proper
precautions such as adequate electrical
grounding of equipment and not
having flammable liquids in the area
should be taken.
For further information, refer to
NFPA 77-1993, Recommended Practice on
Static Electricity.7
Precautions with Krypton-85 Gas
Krypton-85 gas is used in leak test
pressurization systems in concentrations
near 0.01 percent in nitrogen or air. The
112
Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
PART 3. Flammable Liquids and Vapors
Definition of Terms
Characterizing Flammable
Liquids and Vapors
Flammable liquids are usually subdivided
into classes. As defined by the National
Fire Protection Association, a flammable
liquid is any liquid having a flash point
below 60 °C (140 °F) and having a vapor
pressure not exceeding 275 kPa absolute
(40 lbf·in.–2) at 37.8 °C (100 °F).
Combustible liquids are those with
flash points in the range of 60 to 93 °C
(140 to 200 °F). Although they do not
ignite as easily as flammable liquids, they
can ignite under certain circumstances
and so must be handled with caution. The
more common flammable and
combustible liquids are various
hydrocarbons, alcohols and their
byproducts. They are chemical
combinations of hydrogen and carbon;
the combination may also contain
oxygen, nitrogen, sulfur and other
elements.
Factors Influencing Hazards of
Flammable Liquids
Flammable liquids vaporize and form
flammable mixtures when they are in
open containers, when leaks or spills
occur or when the flammable liquids are
heated. The degree of danger depends on
the following: (1) the flash point of the
liquid, (2) the concentration of vapors in
the air (whether the mixture of vapor and
air is in the flammable range) and (3) the
possibility of an ignition source at or
above a temperature sufficient to cause
the mixture to burst into flame.
atmosphere. In both cases, the fluids
should be enclosed wherever feasible.
When the fluid is exposed to air for a
specific operation, it should again be
covered or enclosed as soon as possible.
Flash Point of a Flammable Liquid
The flash point of a liquid is the lowest
temperature at which it gives off enough
vapor to form flammable mixtures with
air and to produce a flame when a source
of ignition is brought close to the surface.
Other properties are factors in
determining the hazards of flammable
liquids, but the flash point is the principal
factor. The relative hazard increases as the
flash point is lowered. The significance of
this property becomes more apparent
when liquids of different flash points are
compared.
Examples of Flash Points of common
Liquid Fuels
Kerosene and number 1 fuel oil have flash
points of about 43 to 74 °C (110 to
165 °F) but ASTM D 396, Specification for
Fuel Oils,8 will permit a flash point as low
as 38 °C (100 °F) for number 1 fuel oil. At
ordinary room temperatures of 22 °C
(72 °F), these oils do not give off
dangerous quantities of vapor. On the
other hand, gasoline gives off vapor at a
rate sufficient to form a flammable
mixture with air at temperature as low as
–45 °C (–50 °F).
Any flammable liquid, when heated to
a temperature above its flash point, can
produce vapors in sufficient quantity to
produce an explosive mixture in the air.
For example, when heated, heavy fuel oil
may produce flammable vapors just as
readily as gasoline does at –20 °C (–4 °F).
Precautions for Flammable
Liquids
Autoignition Temperature
In the handling and use of flammable
liquids, exposure of large liquid surfaces
to air should be prevented. It is not the
liquids themselves that burn or explode,
but rather the vapor-and-air mixture
formed when liquids evaporate. Therefore,
flammable liquids should be handled and
stored in closed containers. Low flash
liquids in use should be covered or
enclosed to avoid evaporation into the
Autoignition temperature is the lowest
temperature at which a flammable gas or
vapor-and-air mixture will ignite under
defined conditions without an external
source of ignition. Flammable vapors and
gases in oxygen will spontaneously ignite
at a lower temperature than in air and
their autoignition temperature may be
influenced by the presence of catalytic
substances.
Safety Aspects of Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
113
Flammability Limits of Vapor
Concentrations
Flammable liquids have a minimum
concentration of vapor in air below which
propagation of flame does not occur on
contact with a source of ignition. There is
also a maximum proportion of vapor or
gas in air above which propagation of
flame does not occur.
The extremes of vapor or gas
concentration with air which, if ignited,
will just propagate flame, are known as
the lower and upper flammable limits.
These are usually expressed in terms of
percentage by volume or weight of gas or
vapor in air. These limits are also
commonly referred to as, respectively, the
lower and upper explosive limits.
A mixture with less than about 1.0
percent by weight of gasoline vapor is too
lean and propagation of flame will not
occur on contact with a source of
ignition. Similarly, if there is more than
about 8 percent of gasoline vapor, the
mixture will be too rich. Other gases such
as hydrogen, acetylene and ethylene have
a wider range of flammable limits.
Flammability Ranges (Explosive
Range)
Flammable range is the difference
between the lower and upper flammable
limits, expressed in terms of percentage by
volume of vapor or gas in air. It is also
often referred to as the explosive range.
For example, the limits of the flammable
range of gasoline are generally taken as
1.4 to 7.6 percent, which is relatively
narrow. Thus, a mixture of 1.4 percent
gasoline vapor and 98.6 percent air is
flammable, as are all the intermediate
mixtures up to and including 7.6 percent
gasoline vapor and 92.4 percent air. The
range is the difference between these
limits, or 6.2 percent.
Effects of Diffusion Rate,
Vapor Pressure and
Volatility
Rate of diffusion is the tendency of one gas
or vapor to disperse into or mix with
another gas or vapor. This rate depends
on the density of the vapor or gas as
compared with that of air. Whether a
vapor or gas is lighter or heavier than air
determines to a large extent the means of
solving ventilation problems.
Vapor pressure is the partial pressure (in
kilopascal or in lbf·in.–2) exerted by the
vapor of a volatile liquid, when in
equilibrium with the surface of the liquid,
as determined by standard ASTM D 323,
Test Method for Vapor Pressure of Petroleum
Products (Reid Method).9
114
Leak Testing
Volatility is the tendency or ability of a
liquid to vaporize. Such liquids as alcohol
and gasoline, because of their well known
tendency to evaporate rapidly, are called
volatile liquids.
Boiling Points of
Flammable Liquids
The boiling point of a liquid is that
temperature at which the vapor pressure
of the liquid equals the atmospheric
pressure. Increasing the liquid
temperature causes vapor to be given off
more readily. Liquids with low boiling
points generally volatilize more readily
than those with higher boiling points.
However, there is not consistent
relationship between boiling point and
evaporation rate.
Definitions for Vapor Volume and
Evaporation Rate
Vapor volume is the number of liters of
solvent vapor formed by evaporation of
1.0 L of liquid at standard temperature
(20 °C). In English units, the vapor
volume is the number of cubic feet of
solvent vapor formed by the evaporation
of 1 gal (imperial or United States gallon),
of a liquid at 68 °F. One can always find
vapor volume by using the mole (an
amount of gas or liquid whose weight in
grams equals its molecular weight.) This
number of grams, equal to the molecular
weight of the substance (at 0 °C and
101.3 kPa), evaporates to 22.4 L at
standard temperature and pressure.
Evaporation rate is the ratio of time
required to evaporate a measured volume
of liquid to the time required to evaporate
the same volume of a reference liquid
under ideal test conditions. The higher
the ratio, the slower the evaporation rate.
Containers for Flammable
Liquids
Portable containers should be provided
with flame arresters installed in the vent
or opening. If a number of different
flammable liquids are handled, safety cans
should have distinct stripes, or
identification lettering should be placed
on them so as to reserve certain cans for
their respective liquids and to help reduce
the chance of the liquids being mixed.
Safety can caps should be regularly
inspected for proper operation and
sealing.
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
Restriction of Smoking and
Lighters in Flammable
Material Areas
Smoking and carrying of lighters, strikeanywhere matches and other spark
producing devices should be prohibited in
buildings or areas where flammable
liquids are stored, handled or used. The
extent of the restricted area will depend
on the type of products handled, the
design of the building design, local
conditions and compliance with local,
state and federal regulations for
flammable material areas.
Safety Aspects of Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
115
PART 4. Electrical and Lighting Hazards
Hazards of Static Electricity
with Flammable Materials
Static electricity is an accumulation of
motionless charges generated by the
contact and separation of dissimilar
materials. For example, static electricity is
generated when a fluid flows through a
pipe or from an orifice into a vessel and
may set up high voltages. The principal
hazards created by static electricity are
those of fire and explosion caused by
spark discharges occurring in the presence
of flammable or explosive vapors, gases or
dust. A spark between two bodies occurs
when there is no good electrical
conductive path between them. Hence,
grounding and bonding of flammable
liquid containers is necessary to prevent
static electricity from causing a spark.
Avoidance of Sources of
Ignition of Flammable
Gases and Vapors
When using potentially explosive gases,
the test area should be free from obvious
sources of ignition. Smoking should be
prohibited and signs should be posted to
warn of the hazards. Electrical equipment
may also present a problem. If there is a
possibility that, in the event of leakage,
such equipment will be in an explosive
environment, then either the equipment
should be repositioned outside the danger
area or specifically chosen equipment
should be used.
Although hydrogen presents the most
severe risk, the above precautions are also
relevant if other flammable tracer gases
are used. When large components are
tested, or when large volumes of
hydrogen are used, it may be advisable to
provide monitors that give a continuous
indication of the hydrogen and air
content in the test area. Intrinsically safe
detectors are available. Gas monitoring
may also be advisable when high vacuum
vessels are being chemically cleaned
before evacuation. Cleaning techniques
often include washing with benzene,
acetone or alcohol. The interior of the
vessel as well as the environment may
contain an explosive mixture. Extreme
precautions should be taken when using
these materials.
116
Leak Testing
A discharge of static electricity is a
possible cause of ignition, so all metal
parts likely to become charged should be
grounded. When testing with gases such
as hydrogen, it would also be sensible for
personnel to avoid wearing clothing that
might produce static charges and for them
to wear shoes with conducting soles.
Another precaution is the use of reduced
sparking or nonsparking tools.
Bonding and Grounding to
Prevent Electric Sparks
A point of great danger from a static spark
is the place where a flammable vapor may
be present in the air, such as at the outlet
of a flammable liquid fill pipe or a
delivery hose nozzle. Static spark ignition
sources are prevented by bonding or
grounding or both so they have the same
static voltage or potential.
The terms bonding and grounding
often have been used interchangeably
because of poor understanding of the
distinct functions indicated. Bonding is
done to eliminate a difference in potential
between objects. The purpose of
grounding is to eliminate a difference in
potential between an object and ground.
Bonding and grounding are effectively
applied only to conductive bodies. The
human body is a conductive body that
may differ in potential from ground or
other bodies, so that it may also serve as a
source of spark ignition.
Although bonding will eliminate a
difference in potential between the
objects that are bonded, it will not
eliminate a difference in potential
between these objects and the earth
unless one of the objects possesses an
adequate conductive path to earth.
Therefore, bonding will not eliminate the
static charge but will only equalize the
potential between the objects bonded.
Electrical Power Hazards
Electricity as a source of power is, in some
ways, less hazardous than steam or other
energy sources. However, failure to take
suitable precautions in its use creates
conditions that are certain to result in
bodily harm or property damage or both.
Although there have been advances in the
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
control of electrical hazards, industry still
has many injuries and fatalities from
preventable causes. Machine tools can,
with minimum expense and difficulty, be
arranged for maximum safety and
efficiency. There are, however, certain
hazards in the installation, maintenance
and use of electric wiring and equipment.
Control of most of these hazards is
neither difficult nor expensive, but
ignoring or neglecting them may lead to
serious accident.
Electrical Injury and Fatal Levels of
Body Current
Current flow is the factor that causes
injury in electric shock. The severity of
electric shock injury is determined by the
amount of current flow through the
victim. Experimental and field data from
authoritative sources indicate that, in
general, an alternating current of 0.1 A at
commercial frequency (60 Hz) may be
fatal if it passes through the vital organs.
Similarly, it is estimated that a current
value of 0.02 A is the limit at which an
individual can still release himself from
an object held by the hand. Such current
flow may readily result from body contact
with low voltage sources of ordinary
lighting or power circuits.
Limiting Current Flow to
Human Body
Because current flow depends on voltage
and resistance, these factors are
important. Other factors affecting the
amount of injury are the parts of the body
involved, the duration of current flow
through the victim and the frequency
with alternating current. Resistance to
current flow is mainly to be found in the
skin surface. Callous or dry skin has a
fairly high resistance, but a sharp decrease
in resistance takes place when the skin is
moist. Once the skin resistance is broken
down, the current flows readily through
the blood and body tissues. Grounding
conditions often determine resistance to
current flow from the human body to
earth or grounded structures.
Whatever protection is offered by skin
resistance decreases rapidly with increase
in voltage. High voltage alternating
current at 60 Hz causes violent muscular
contraction, often so severe that the
victim is thrown clear of the circuit.
Although low voltage also results in
muscular contraction, the effect is not so
violent. The fact, however, that low
voltage often prevents the victim from
freeing himself from the circuit makes
exposure to it dangerous.
Effects of Electric Current
on Human Body
Death or injury by electric shock may
result from the following effects of current
on the body.
1. Electric current may cause contraction
of the chest muscles, which may
interfere with breathing to such an
extent that death will result from
asphyxiation when the exposure is
prolonged.
2. Electric current may cause temporary
paralysis of the nerve center, which
may result in failure of respiration, a
condition that often continues until
long after the victim is freed from the
circuit.
3. Electric current may interfere with
normal rhythm of the heart, causing
ventricular fibrillation. In this
condition, the fibers of the heart
muscles, instead of contracting in a
coordinated manner, contract
separately and at different times.
Blood circulation ceases and death
ensues, because apparently the heart
cannot spontaneously recover from
this condition. It has been estimated
that 0.1 A flowing through the body
cavity (chest) is sufficient to cause
ventricular fibrillation.
4. Electric current may suspend heart
action by muscular contraction (on
contact with heavy current). In this
case, the heart may resume its normal
rhythm when the victim is freed from
the circuit.
5. Electric current may cause
hemorrhages; destruction of nerves,
muscles or other tissues; or extensive
skin burn from heat due to heavy
current or electric arcs.
In general, the longer the current flows
through the body, the more serious may
be the result. Considerable current is
likely to flow from high voltage sources
and in general only very short exposure
can be tolerated if the victim is to be
revived.
Injuries from electric shock are less
severe when the current does not pass
through or near nerve centers and vital
organs. In most electric accidents in
industry, the current flows from hands to
feet. Because such a path involves both
the heart and the lungs, results are usually
serious.
Treatment of Victims of
Electric Shock
Statistics indicate that only a small
percentage of those who recover from
electric shock show permanent disability.
Safety Aspects of Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
117
In many cases, the victim may be saved
by prompt application of
cardiopulmonary resuscitation because a
common result in electrical accidents is
failure of that part of the nervous system
that controls breathing. Therefore, it is
essential that persons working with
electrical power equipment be instructed
in the modern technique of
mouth-to-mouth or mouth-to-nose
resuscitation and cardiopulmonary
resuscitation as developed by the
American Heart Association. Immediate
treatment should be applied to victims of
electric shock and should be continued
until they revive or until death is
diagnosed by a physician or until rigor
mortis sets in.
Hazards of Electric Arcs
Another type of injury is burns from
electric arc flashes or from human contact
with energized electric power equipment.
Such burns are usually deep and slow to
heal and may involve large areas of the
body. Even welding arcs are also sources
of arc flashes. Hot weld metal, welding
slag and electrode stub ends can produce
severe burns if touched. Side shielded
safety glasses, glasses that do not transmit
ultraviolet radiation and proper use of
welding helmets all help avoid welding
arc flash injuries to the eye.
Hazards of Electrical Extension Cords
Extension cords should be of a type listed
by the Underwriter’s Laboratories and
should be labeled to show that they meet
all requirements of the National Electrical
Code.10 They should be inspected
regularly. Kinking or excessive bending of
the cord should be avoided to prevent the
wire strands from breaking. Broken
strands may pierce the insulated covering
and become a shock or short circuit
hazard. Old insulation on extension cords
often becomes brittle and creates a shock
or short circuit hazard. Ordinary twisted
lamp cord should never be used for
extension cords or lamps in vessels or on
damp or metallic floors and should never
be used where it will be exposed to
mechanical wear.
Cord for use with portable power tools
and equipment is made in several grades,
each of which is designed for a specific
type of service. Rubber sheath cord should
be used with portable electric tools and
with extension lamps in vessels or other
grounded enclosures. Special types of
synthetic rubber or plastic covering
should be considered when the cord is to
be used in areas where it may come in
contact with oils or solvents. Double
insulated electrical tools should be
selected for maximum safety.
118
Leak Testing
Explosion-Proof Electrical
Fittings
When using potentially explosive gases,
the test area should be free from obvious
sources of ignition. Smoking should be
prohibited and signs should be posted to
warn passersby of the hazards. Electrical
equipment may also present a problem. If
there is a possibility that, in the event of
leakage, such equipment will be in an
explosive environment, then either the
equipment should be repositioned outside
the danger area, or else specifically
chosen, safe, explosion-proof equipment
should be used.
Standard electrical fittings, considered
safe for ordinary application, are
obviously unfit for installation in
locations where flammable gases and
vapors or other easily ignitable flammable
materials are present. Sparks and electric
arcs originating within electrical switches
and fittings have been the igniting
medium in costly fires and explosions.
Selection of Electrical Fittings for
Hazardous Locations
Before fittings are selected for a hazardous
location, it is necessary to determine the
exact nature of the flammable materials
present. For instance, an electrical fitting,
found by test to be safe for installation in
an atmosphere of combustible dust, may
be unsafe for operation in an atmosphere
containing flammable vapors or gases.
It is impossible to prevent highly
flammable gases from entering the
interior of either an explosion resistant or
an ordinary wiring system. They will
eventually enter the entire line through
the joints and through the breathing of
the conduit system caused by temperature
changes. Furthermore, gaseous vapors will
fill every crevice whenever covers are
removed. For these reasons, it is
impossible to provide an entirely vaporproof switch unit or to regulate
temperatures or keep the air free from
flammable gases inside the electrical
fittings.
To protect that area classified as a
hazardous location, it is necessary to have
positive confinement of the arc, heat and
explosion within the internal limits of
explosion-proof fittings. These fittings are
constructed to completely imprison the
dangerous arcing, intense heat and
subsequent explosion so that the gas
laden air outside does not become ignited.
Protective Enclosures for Electrical
Apparatus
A useful substitute for explosion-proof
equipment is to enclose nonexplosion-
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
proof apparatus in metal boxes and pass a
stream of nitrogen or even air into the
box to maintain it slightly above
atmospheric pressure. However, it should
be kept in mind that all equipment can be
hazardous and should only be used with
due regard to the hazards involved. Put
only that equipment in the test area that
must be there. Where possible, use air
operated equipment instead of electrical
equipment.
It is necessary that electrical equipment
be explosion-proof throughout the entire
building and not solely in the test area, in
cases where explosive vapors may travel
to other parts of the building should a
leak occur. Under some circumstances, the
test area can be sealed to prevent escape
of vapors to other areas.
sufficient light for general safety and for
ordinary visual needs.
Light intensity on a surface varies
inversely with the square of the distance
between the surface and a small area
source of light. A source 3 m (about 10 ft)
above a surface would give four times
more light to the work area than would
the same source 6 m (about 20 ft) high.
Where visual needs are more critical,
additional lighting can thus be provided
by fixtures placed fairly close to the area
needing more light.
Lighting as a Factor in
Industrial Safety
Fluorescent penetrants and leak tracers
require intense illumination with
ultraviolet and near ultraviolet radiation
sources to make test indications visible.
Properly enclosed, shielded and filtered
ultraviolet radiation sources used for
inspection emit radiation in the 320 to
400 nm wavelength range, well above the
more hazardous shorter wavelength
ranges of hard ultraviolet radiation.
Failure to use proper filters and lamp
enclosures could permit such hard
ultraviolet radiation from mercury vapor
arc lamps, welding arcs and fluorescent
tubular ultraviolet lamps to escape. The
following discussion lists hazards and
precautions for control of ultraviolet
radiation and notes its physiological
effects.
The proportion of industrial accidents
attributable to poor lighting has been
estimated to be from 15 to 25 percent.
Good lighting contributes greatly to
safety, as well as increasing efficiency and
morale. Daylight is an ideal type of
illumination. For the most effective use of
daylight, a definite relationship of floor to
window must be maintained. Sudden
transitions from brightly lighted to dim
areas and vice versa are dangerous; the
result is momentary blindness due to the
lag in eye accommodation. Gradations of
light between areas of different intensities
will remedy this difficulty.
Precautions with
Ultraviolet Sources Used
for Inspection with
Fluorescent Leak Tracers
Artificial Lighting
Artificial lighting has become so accepted
as an element of modern life that its
original supplementary character has been
largely forgotten. Artificial lighting has
become the major source of illumination
because natural light is undependable,
especially in the winter when work
schedules do not coincide with daylight.
For continuous shift operation, artificial
light is essential. For other types of
operation, it must be relied on from 20 to
50 percent of the total working hours,
excluding overtime work or night work.
General Lighting
General lighting is the base or minimum
amount of light required. It has been
defined as uniform distribution of light to
produce equivalent seeing conditions
throughout an interior. Localized general
lighting sources usually are arranged 3 m
(about 10 ft) or more above the work to
prevent too great a contrast in brightness
between the more highly lighted work
area and the adjacent areas and to provide
Effects of Hard Ultraviolet
Radiation
Hard (short wavelength) ultraviolet
radiation has long been known to
produce physical, chemical and
physiological effects, so some evaluation
of these effects and the degree of hazard
involved in ultraviolet radiation is in
order.
Physically, ultraviolet radiation is the
portion of the electromagnetic spectrum
with wavelengths between those of visible
light and X-rays. Therefore, as might be
expected, the long wave portions behave
very much like visible light and the short
wave portions have some of the properties
of X-rays. The middle ranges have
properties of their own that are not
common to other portions of the
spectrum.
Safety Aspects of Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
119
Filters for Transmission or
Absorption of Ultraviolet
Radiation
Ultraviolet radiation can be transmitted,
absorbed and refracted or bent just like
visible light, although usually by
substances other than those normally
used for visible light. For instance,
ordinary window glass transmits quite
well in the longer wavelengths, but
becomes opaque to wavelengths shorter
than 310 nm. Thus, it will transmit
ultraviolet radiation but absorb the
shorter, more harmful wavelengths.
Therefore, ordinary glass is a good
protective shield against hard ultraviolet
radiation.
A number of suitable filters and glass
types will remove all ultraviolet radiation
while permitting visible light to pass. In
cases where short wave ultraviolet
radiation must be transmitted, special
glasses are available. Some glass will
transmit wave lengths as short as 280 nm
and other glass to 230 nm. Below this
point, quartz, particularly in the
crystalline form, must be used.
Reflection of Ultraviolet
Radiation
Ultraviolet can also be reflected, but often
by materials different from those used to
reflect visible light. Most white metals
reflect ultraviolet radiation although not
as strongly as they reflect visible light.
Silver is an exception, reflecting to about
360 nm, with absorbing shorter
wavelengths. Aluminum and polished
iron are good ultraviolet reflectors. Some
white pigments such as magnesium oxide,
aluminum oxide and calcium carbonate
are good reflectors, whereas others such as
titanium dioxide and zinc oxide are poor
ultraviolet reflectors. Dark visible colors,
particularly greens, browns and reds, are
usually poor reflectors and good absorbers
of ultraviolet. These factors should be kept
in mind when designing ultraviolet
radiation inspection booths.11
Chemical Reactions Excited
by Hard Ultraviolet
Radiation
Ultraviolet radiation is also chemically
active and accelerates many reactions. Of
particular importance are oxidation and
molecular breakdown. Oxidation is a
primary cause for the breakdown of paint
vehicles and the fading of dyes and other
colorants. Powerful ultraviolet will also
120
Leak Testing
bread down or otherwise alter many
molecules even without the presence of
oxygen, so oxidation is not the only
chemical action it produces. Ultraviolet
radiation, being at the long wave end of
the range, is probably only slightly more
chemically active than visible light.
However, as with visible light, long
exposure to high intensity ultraviolet
radiation can be expected to have its
effect.
Chemical Reactions Caused
by Ozone
Hard (short wave) ultraviolet radiation
also produces ozone, which itself is a
strong oxidant. Ozone (O3) is a bluish gas
with a characteristic pungent odor and is
found naturally in the atmosphere as a
result of the action of solar radiation and
lightning in electrical storms. It is also
formed in corona discharges around high
voltage conductors and is generated by
X-ray and ultraviolet radiation, electric
arcs (including welding arcs), mercury
vapor lamps and linear accelerators.
Physiological Effects of
Hard Ultraviolet Radiation
Physiologically, ultraviolet radiation can
produce a variety of effects, depending
strongly on the wavelength. Short wave
ultraviolet radiation, as previously stated,
produces ozone. Ozone is a very toxic
compound that may cause death due to
lung congestion and edema. Its maximum
allowable concentration is 0.1 part per
million (0.2 mg·m–3). Ozone is produced
essentially at wavelengths below 260 nm.
The properly filtered mercury arc
ultraviolet radiation sources used with
fluorescent leak tracers do not produce
ozone.
Ultraviolet radiation also has a
germicidal effect and is used for
sterilization. This effect reaches a
maximum at 260 nm and falls off rapidly
to nearly zero at 320 nm. The action is
effective on almost all bacteria as well as
some fungi and molds. Thus, ultraviolet
radiation is a very useful tool for
disinfecting surfaces as well as room air
while it passes through enclosed
ventilating systems. Sterilizing lamps
should not be placed where human eyes
or skin can be exposed to their radiation.
Skin Inflammation Caused by
Ultraviolet Radiation
Another well known effect of ultraviolet
radiation is the production of erythema or
skin inflammation, commonly known as
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
sunburn. This effect is produced strongly
by certain wavelengths and not at all by
others, as shown in Fig. 1. Thus, the short
germicidal wavelengths produce
considerable inflammation, whereas
certain middle wavelengths are relatively
ineffective. For those using ultraviolet
radiation for inspection, the important
fact is that there is essentially no
erythemal effect above 320 nm. Because
the light used for inspection is essentially
365 nm, inspectors do not become
sunburned from their work with properly
filtered and shielded black lights.
One of the serious concerns about
possible effects of any radiation is its
tendency to produce cancer. The United
States government in its role as a
consumer protection agency has
conducted studies on carcinogenic and
other health hazards of ultraviolet
radiation. This work is summarized in the
document Criteria for a Recommended
Standard for Occupational Exposure to
Ultraviolet Radiation,12 which concludes
that cancer can be produced by long
exposure to sunlight rich in midrange
ultraviolet. However, the cancer
producing effect is directly proportional
to the erythemal or sunburn effect.
Therefore, the ultraviolet radiation used
for inspection purposes is not a probable
cause of cancer.
FIGURE 1. Standard curve for erythemal effectiveness of
various wavelengths of hard ultraviolet radiation. Note that
radiation used in inspection with fluorescent leak tracers lies
in the range of 360 nm and above and does not have
significant hazardous effects.
Eye Irritation Caused by Ultraviolet
Radiation
Eye irritation is one further physiological
effect due to ultraviolet radiation. There
are two types of irritation. The first is a
bluish haze noted when the eyes are
exposed to ultraviolet, particularly of the
longer wavelengths. This is irritating,
causing headaches and, in extreme cases,
nausea but is otherwise not harmful. It is
caused by fluorescence of certain portions
of the eye when exposed to ultraviolet
radiation.
The second type of irritation is
photokeratitis followed by conjunctivitis.
This is essentially snow blindness. It
includes a feeling of sand in the eyes,
allergy to light, tear formation and finally
blindness. These symptoms usually begin
6 to 12 h after exposure and last from 6 to
24 h, with all symptoms disappearing in
48 h. There is not cumulative effect but,
on the other hand, no tolerance is
developed from repeated exposure as is
the case with sunburn. This effect is
caused only by the wavelengths shorter
than 310 nm and so should be no
problem in inspection operations as long
as the light is passed through the normal
filters.
Protective Glasses to Shield Eyes from
Ultraviolet Radiation
Should eye irritation be a problem, yellow
tinted eye glasses will offer complete
protection, particularly if equipped with
side shields. Such glasses are sold as
shooter’s glasses and can be provided by
local oculists.
1.0
Recommended Limits for
Personnel Exposure to
Ultraviolet Radiation
0.9
0.8
In Criteria for a Recommended Standard for
Occupational Exposure to Ultraviolet
Radiation,12 recommended limits for
personnel exposure to ultraviolet
radiation in the 314 to 400 nm range
were listed as 1.0 mW·cm–2 for exposures
exceeding 1000 s and 100 mW·cm–2 for
exposures under 1000 s (about 16 min).
The OSHA Environmental Standard
was 10 mW·cm–2 over any 1 h period.
Relative effectiveness
0.7
0.6
0.5
0.4
0.3
0.2
Recommended Limits for
Exposure to Krypton-85
Gas
0.1
250
260
270
280
290
Wavelength (nm)
300
310
Although regulations in some of the
United States vary in specific exposure
limits, the Code of Federal Regulations,
Title 10, Part 20,1 sets standards for the
Safety Aspects of Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
121
allowable cumulative annual exposure:
1 mSv (100 mrem) for the general public,
50 mSv (5 rem) for the whole body,
150 mSv (15 rem) for the lens of the eye,
0.5 Sv (50 rem) for the skin and
extremities and 5 mSv (500 mrem) for an
embryo or fetus.
The prevailing industrial philosophy is
that any unnecessary exposure should be
prevented. The gamma radiation from
krypton-85 has a 514 keV energy and
represents 0.46 percent of the emission
from krypton-85 gas. This is considered to
be a very week photon with potential for
very little tissue damage. The beta particle
emitted by krypton-85 is quite weak and,
when the gas is leaked into a hermetic
device, the beta particles rarely can
penetrate the walls of the device. In actual
leak testing, the krypton-85 gas that has
leaked into the device is measured by
detecting the total radiation seen through
the walls of the device using highly
sensitive scintillation detectors.
122
Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
PART 5. Safety Precautions with Leak Testing
Tracer Gases
Tracer Gas Hazards in Leak
Testing
Tracer gas safety aspects such as
flammability, asphyxiation or specific
physiological effects as well as the
possibility of pressure vessel explosions
must be considered. So long as the
nondestructive test engineer and the leak
test technician are aware of these
considerations from the start, it is possible
to leak test a vessel with minimum
inconvenience or danger.
Most tracer gases are not toxic.
However, if a question exists about the
toxicity of any particular gas, a competent
authority should be consulted. Many
tracer gases will not support human life. If
such tracer gases replace oxygen in a
vessel, this vessel cannot be entered
without proper ventilation. In this case,
proper ventilation consists of a gas mask
that contains its own air-oxygen gas
supply.
The oxygen required for breathing may
be accidentally removed from an area. For
example, if one of the halogenated
hydrocarbons is used as a tracer gas, it
may stagnate and settle to the lowest area.
If a technician is attempting to use a
detector probe in this low area, the tracer
gas that settles may eventually displace
enough of the air to produce
asphyxiation. To avoid this condition,
adequate ventilation must be provided.
However, this ventilation must be
performed carefully. If the tracer gas is
removed too rapidly from the place where
it is escaping from the vessel, leakage
location may be difficult.
To aid in a better understanding of the
safety aspects, the following data are
presented below for several tracer gases
that may be used. In addition,
information is given on the availability of
personnel protection indicators and area
contamination monitors that can provide
warning indications of dangerous
accumulations of toxic gases or vapors.
Personnel Protection
Badges to Warn of
Excessive Exposure to
Toxic Gases
Personnel protection indicators (PPIs) are
plastic badges with pocket clips that have
sensors that react chemically with
concentrations of various gases or vapors
used as tracers in leak testing. They
provide forewarning of excessive exposure
to the toxic substances by means of color
changes, as listed in Table 5. These
personnel protection indicators are
sensitive to the accumulated personal
exposure of the badge wearer to the
concentration of gas in the leak testing
area. The Occupational Safety and Health
Administration of the United States
defines the critical exposure period to be
an 8 h shift. A color change of the
protective badge at any time during an
8 h shift indicates that the badge wearer
has received his or her maximum safe
exposure. Table 5 lists the concentrations
of toxic gas or vapor in air, which are
designated as the critical accumulations.
Also listed in Table 5 are the color
changes that occur on exposure of
personnel protection badges to the
specific tracer gases for which they are
sensitive.
Although the personnel protection
indicator badges are normally worn on
TABLE 5. Selection guide for personnel
protection indicators for toxic gases and
vapors accumulating in leak testing
areas. Data apply to both personnel
protection and area contamination
monitors.
Toxic
Substance
Ammonia
Carbon monoxide
Chlorine
Hydrazine
Hydrogen sulfide
Nitrogen dioxide
Ozone
Warning
Concentration Color
(µL·L–1)
Change
15
50
2
5
5
1
0.1
Brown to white
White to Black
White to yellow
White to yellow
White to brown
White to yellow
White to brown
Safety Aspects of Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
123
breast pockets so supervisory personnel
and coworkers can easily see the status of
the indicator, a person working alone can
monitor his or her own status more easily
by clipping the badge to his or her belt.
Replacement color change buttons are
available to be inserted into these badges
because the color changes occurring on
exposure are permanent. Continuous use
of suitable personnel protection indicators
would be appropriate during leak testing
operations. In addition, such leak testing
areas can be monitored by area
contamination monitors, as described
next.
Contamination Monitoring
of Excessive Accumulations
of Toxic Gases
Area contamination monitors (ACMs) for
atmospheric accumulations of gases and
vapors such as ammonia, chlorine,
hydrazine, hydrogen sulfide, nitrogen
dioxide or ozone are self-adhesive filter
papers that chemically react to
concentrations of various gases or vapors.
Indicating by means of color changes
listed in Table 6, these area monitoring
indicators are normally mounted on walls
or bulkheads that are easily seen by
supervisory personnel and by leak testing
workers. Ideally, the monitors should be
placed opposite an entrance door with a
window (within buildings) or at locations
where they are visible prior to entry in
open areas, so that personnel can see their
indications and do not enter any
contaminated areas unnecessarily.
By contrast, the area contamination
monitors for carbon monoxide is a
triangular wall mounting plaque. The
propane monitor is a vial of crystals. Both
of these monitoring indicators change
color, as indicated in Table 6, when
excessive accumulations of the specific
toxic gas are present.
Each of the contamination monitors
listed in Table 6 indicates the accumulated
exposure to the specific gas to which the
work area has been exposed during the
8 h measurement period set by the
Occupational Safety and Health
Administration. A color change at any
time during this 8 h interval indicates
that anyone in the area is being exposed
to a toxic gas concentration in excess of
the safe maximum.
Portable Electronic
Instrument for Locating
Small Combustible or Toxic
Gas Leaks
Figure 2 shows a portable, hand held
electronic sensing instrument with
pointing indicator; the instrument is used
both for personnel protection and as a
tracer gas detector in leak testing. It
detects all combustible and many
noncombustible toxic gases and vapors,
including the following: acetone, alcohol,
ammonia, benzene, butane, carbon
monoxide, carbon tetrachloride, ethane,
ethylene oxide, gasoline, hydrogen,
FIGURE 2. Portable personnel protection monitor and
detector for leak testing of certain combustible or toxic
gases.
TABLE 6. Selection guide for area
contamination monitors for toxic gases
and vapors accumulating in leak testing
areas.
Toxic
Substance
Critical
Concentration
(µL·L–1)
Ammonia
15
Carbon monoxide 50
Chlorine
2
Hydrazine
5
Hydrogen sulfide
5
Nitrogen dioxide
1
Ozone
0.1
Propane
0.001
124
Leak Testing
Color
Change
Brown to white
White to Black
White to yellow
White to yellow
White to brown
White to yellow
White to brown
Purple to yellow
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
turpentine, hydrogen sulfide, liquid
propane gas, methane, methyl ethyl
ketone, naphtha, natural gas, propane,
steam, sulfur dioxide, toluene,
trichloroethylene and xylene.
This instrument does not detect carbon
dioxide. It is a low cost, simple leak tracer
designed to locate small leaks. A flexible
1 m (3 ft) extension hose can be used to
sniff leaks in less accessible locations
behind pipes or around complex pipe
connections. Its use is often more
convenient than using bubble tests and
the small battery operated hand held
detection and indicating instrument is
often more feasible than larger electronic
instruments requiring connections to
alternating current power outlets. Its
sensor is reported to detect 50 µL·L–1 of
gas or vapor contaminant in atmospheric
air and is designed for over 50 000
exposures to gases.
Precautions with Ammonia
Gas
Ammonia (NH3) is used as a tracer gas for
many chemical indicator leak tests. At
room temperature and atmospheric
pressure, ammonia is a colorless, alkaline
gas having a pungent odor, which
provides ample warning of its presence.
Ammonia gas is irritating to the eyes and
to moist skin. However, concentrations of
ammonia gas in air in the concentration
range below 50 µL·L–1, although not
harmful, are a considerable nuisance, so
that people tend to avoid them. It is
therefore unlikely that an individual
would unknowingly become overexposed
to ammonia gas.
Physiological Effects of Ammonia Gas
Table 7 lists the physiological effects of
various concentrations of ammonia. The
corrosive action of high concentrations
(above 700 µL·L–1) can cause extensive
injuries to the eyes, including severe
irritation, hemorrhaging and swollen lids.
If not treated immediately, partial or total
loss of sight may result. The mucous
lining of the mouth, throat, nose and
lungs is particularly sensitive to ammonia
attack.
liquid ammonia from the skin surface can
cause frostbite. Anyone working with
liquid ammonia must wear rubber gloves,
chemical protection clothing and goggles
and a rubber or plastic apron.
Hazards of Explosion or Ignition with
Ammonia
Ammonia cylinders should never be
directly heated by steam, direct electric
coils or flames. Uncontrolled heating of a
cylinder can cause the liquid to expand to
a point where dangerous pressures will be
developed. Heating is done in a
thermostatically controlled water or oil
bath. In no case should the temperature
be allowed to exceed 50 °C (120 °F).
Ammonia represents a possible
flammability hazard. A mixture of air and
ammonia containing from 15 to
28 percent ammonia by volume will
ignite when sparked or exposed to
temperatures exceeding 50 °C (120 °F).
Therefore, flames and sparks should not
be allowed in the area where ammonia is
being used.
As another noteworthy consideration,
ammonia can combine with mercury to
form explosive compounds. Therefore,
instruments containing mercury (such as
manometers) should not be used where
they will be exposed to ammonia.
Precautions with Argon
Gas
On some occasions, argon (Ar) is used as a
leak tracer gas. It is the most abundant
member of the rare gas family, which
consists of helium, neon, argon, krypton
and xenon. All of these gases are
monatomic and are characterized by their
extreme chemical inactivity. Argon, a
TABLE 7. Physiological effects of various concentrations of
ammonia gas (NH3).
Atmospheric
Concentration
(µL·L–1)
20
40
100
Precautions with
Anhydrous Liquid
Ammonia
Contact with anhydrous liquid ammonia
is intensely irritating to the mucous
membranes, eyes and skin. Contact with
the skin will produce severe burns and the
freezing effect due to rapid evaporation of
400
700
1700
5000
Physiological Effects
First perceptible odor.
A few individuals may suffer slight eye
irritation.
Noticeable irritation of eyes and nasal passages
after few minutes’ exposure.
Severe irritation of the throat, nasal passage
and upper respiratory tract.
Severe eye irritation. No permanent effect if
the exposure is limited to less than 0.5 h.
Serious coughing, bronchial spasms; less than
0.5 h of exposure may be fatal.
Serious edema, strangulation, asphyxia, fatal
almost immediately.
Safety Aspects of Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
125
colorless, odorless and tasteless gas, is
nontoxic. However, argon can act as a
simple asphyxiant by displacing the
amount of air necessary to support life.
Precautions with Carbon
Dioxide Gas
Carbon dioxide (CO2) is a nonflammable,
colorless, odorless and slightly acid gas
which is about one and one half times as
dense (heavy) as air. The normal
concentration of carbon dioxide in the
atmosphere is 0.03 percent, or 300 µL·L–1.
Gaseous carbon dioxide is not a
chemically active compound as such and
high temperatures are generally required
to promote its chemical reactions.
However, aqueous solutions of carbon
dioxide are acidic and many reactions
occur readily.
When it replaces breathable air, carbon
dioxide acts as a simple asphyxiant.
Because it is heavier than air and does not
diffuse readily, pure carbon dioxide may
collect in confined, unventilated areas or
in lower regions of large vessels. Gaseous
carbon dioxide is the regulator of the
breathing function. An increase in the
amount of carbon dioxide inhaled will
cause an increased rate of breathing. The
body, while exercising, will burn more
oxygen and the product of this
combustion will be higher concentrations
of carbon dioxide. These higher
Characteristics of
Refrigerant-12 Gas
The halogen tracer gas
dichlorodifluoromethane (CCl2F2) was
widely used in the 1980s. This was the
refrigerant-12 gas used in air conditioners.
It is a colorless, nonflammable gas at
normal temperatures and pressures. In
concentrations of less than 20 percent (by
volume), refrigerant-12 is odorless. At
high concentrations, its odor is mild and
somewhat ethereal and similar to that of
carbon tetrachloride. Refrigerant-12 is
readily liquefied and is usually supplied in
steel cylinders as a liquefied gas under its
own vapor pressure of about 480 kPa
(70 lbf·in.–2 gage) at 21 °C (70 °F).
Refrigerant-12 gas has also been known by
several trade names, including Freon® 12.
Its extensive use as a propellant for spray
cans has been discontinued in the United
States. Its manufacture in and its
importation into the United States have
been banned. However, if this gas is
sprayed on very hot metallic surfaces or in
the presence of flames, it can dissociate to
form deadly toxic gases such as phosgene.
Refrigerant-12 gas is practically
nontoxic. It shows no toxic effects in
guinea pigs in concentrations up to at
least 20 percent by volume for 2 h
exposures. In higher concentrations,
refrigerant-12 may produce some
physiological action, caused primarily by
oxygen deficiency. The generally accepted
maximum allowable refrigerant-12
concentration for an 8 h daily exposure of
personnel is 1000 µL·L–1.
TABLE 8. Physiological effects of carbon
dioxide gas in air.
Carbon Dioxide
Gas in Air
(mL·L–1)
1 to 10
20
30
50
Increased
Lung Ventilation
Slight and unnoticeable
50 percent
100 percent
300 percent (breathing
becomes laborious)
concentrations of carbon dioxide produce
higher rates of breathing listed (Table 8).
Concentrations of 10 percent
(100 000 µL·L–1) of carbon dioxide in
breathing air can produce
unconsciousness; concentrations of 10 to
25 percent may cause death with
exposures of several hours. A
concentration of 5 percent may produce
shortness of breath and headache.
Continuous exposure to 1.5 percent
carbon dioxide may cause changes in
some physiological processes.
126
Leak Testing
Precautions with Helium
Gas
Helium (He) is widely used as a tracer gas
in leak testing with the mass spectrometer
leak detector. It is the lightest member of
the rare gas family and is a chemically
inert, colorless, odorless and tasteless gas.
Helium is not toxic but can act as an
asphyxiant by displacing the air necessary
to support life. Because of its low density,
helium tends to rise to the top regions of
closed vessels or enclosures, where it
could lead to asphyxiation of workers at
these elevations.
Characteristics of
Hydrogen Chloride Gas
To some degree, hydrogen chloride (HCl)
has also been used as a tracer gas.
Anhydrous hydrogen chloride is a
colorless, pungent, nonflammable,
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
corrosive gas with a suffocating odor. It is
heavier than air, soluble in water and
fumes strongly in moist air. The aqueous
solution is known as hydrochloric acid (or
muriatic acid) and may contain as much
as 38 percent hydrogen chloride.
Hydrogen chloride is supplied in cylinders
in the form of a gas over a liquid. The
cylinder pressure is about 4.2 MPa
(610 lbf·in.–2 gage) at 21 °C (70 °F). As
long as liquid is present in the cylinder,
this pressure remains fairly constant.
When the liquid phase is exhausted,
cylinder pressure drops rapidly.
Physiological Effects of Hydrogen
Chloride Gas
Hydrogen chloride is a highly toxic gas
that severely irritates the upper respiratory
tract and is corrosive to the eyes, skin and
mucous membranes. It may produce
dermatitis on repeated exposures. Eye
contact may result in reduced vision or
blindness. Ingestion may be fatal.
Hydrogen chloride concentrations of 0.13
to 0.2 percent (1300 to 2000 µL·L–1) in air
are lethal for human beings in exposure
lasting only a few minutes. The maximum
hydrogen chloride concentration that can
be tolerated for exposures of 60 min is in
the range of 0.005 to 0.01 percent (50 to
100 µL·L–1). However, the unpleasant
effects of hydrogen chloride provide
adequate warning, leading to prompt
voluntary withdrawal of personnel from
hydrogen chloride contaminated
atmospheres.
Precautions with Hydrogen Chloride
Gas
Personnel who handle hydrogen chloride
gas must wear protective clothing such as
rubber or plastic aprons, rubber gloves
and suitable gas tight safety goggles.
Appropriate gas masks with cannisters or
supplied air respirators should be provided
when hydrogen chloride vapor
concentrations are excessive. Woolen
outside clothing or other acid-resisting
fabrics are also recommended for
personnel handling hydrogen chloride.
Personal hygiene and showering after
each work shift should be encouraged.
When hydrogen chloride is supplied from
cylinders, users should always shut off
their hydrogen chloride lines from the use
end, closing valves successively backward
to the cylinder.
Dry gaseous hydrogen chloride is
essentially inert to metals and does not
attack the commonly used structural
metals under normal conditions of use
(room temperature and atmospheric
pressure). In the presence of moisture,
however, hydrogen chloride will corrode
most metals (other than silver, platinum
or tantalum). When hydrogen chloride is
used at higher pressures, it is necessary to
use extra-heavy steel pipe throughout. No
galvanized pipe or bronze valves should
be used.
Precautions with Hydrogen
Gas
Hydrogen (H2) is colorless and odorless
and is the lightest gas known. It is
nontoxic but can act as an asphyxiant by
displacing the necessary amount of air
required to support life. Because hydrogen
is much lighter than air, it tends to collect
near the top of closed vessels. Hydrogen,
in combination with air or oxygen, can
explode with great violence. Hydrogen
gas, although relatively inactive at
ambient temperatures, reacts with almost
all the other elements at high
temperatures and is considered to be a
very dangerous tracer gas. For this reason,
hydrogen should be avoided if at all
possible.
When large vessels are tested or when
large volumes of hydrogen are used, it
may be advisable to provide monitoring
equipment that gives a continuous
indication of the hydrogen and air
content in the test area. Intrinsically safe
detectors are available. This precaution
may also be advisable when high vacuum
vessels are in the process of being
chemically cleaned before evacuation
because the vessel interior as well as the
surrounding environment may contain an
explosive mixture.
Precautions with
Radioactive Krypton-85
Gas
Krypton gas is completely chemically
inert and, thereby, forms no chemical
combinations with any material used in
the tested components. Radioactive
krypton-85 tracer gas, due to its chemical
inertness, does not participate in any
metabolic processes in the body if inhaled
or ingested in any way. If accidentally
inhaled for a short time, normal breathing
of noncontaminated air will rapidly
remove the radioactive krypton gas from
the lungs and body tissue into which it
might be diffused. With an adequate
ventilating system, proper gamma ray
shielding of storage tanks and reasonable
care, krypton-85 tracer gas can be handled
with negligible risk to the operators.
Krypton-85 has a radioactive half life
or 10.76 yr. Over 99 percent of the
disintegrations give no gamma rays but
emit beta particles with a maximum
energy of 0.67 MeV. Only 0.7 percent of
Safety Aspects of Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
127
the disintegrations yield 0.514 MeV
gamma rays. The primary usefulness of
krypton-85 for leak testing depends on
this small proportion of gamma emitting
disintegrations, reinforced in some
applications by the emission of low
energy bremsstrahlung or very soft X-rays.
Many industrial hand held portable
survey meters are used to detect the
presence of trace quantities of krypton-85
gas. The high percentage of beta particle
emission allows for detection of
nanocuries of krypton-85 gas in the air.
Additionally, all equipment approved to
handle krypton-85 gas is required to have
air monitors in continuous operation to
detect any airborne krypton-85 gas and
initiate an alarm.
Precautions with Methane
Gas
Methane is sometimes used as leak testing
tracer gas. Natural gas consists primarily
(85 percent) of methane. Methane gas
(CH4) in its pure state is flammable,
colorless, odorless and tasteless and is not
considered toxic. It can act as a simple
asphyxiant where, present in high
concentrations, it displaces the oxygen
necessary to sustain life. As an example,
coal miners frequently breath air
containing 9 percent methane and do not
appear to suffer. When concentration
increases above this point, pressure on the
forehead and eyes is noticed. However,
this pressure disappears again on
breathing fresh air. Methane in mixtures
with air or oxygen burns rapidly. Ignition
leads to explosions similar to many coal
mine explosions. Incomplete combustion
of methane gas may produce carbon
monoxide, a toxic gas.
Precautions with Nitrogen
Gas
Nitrogen (N2) is not often used as a tracer
gas but may be used to backfill vacuum
vessels or may be mixed with a tracer gas
and introduced into a vessel before a
pressure leak test. Nitrogen gas comprises
about 79 percent by volume of the air. It
will not burn and will not support
combustion. It is nontoxic; however,
nitrogen can act as an asphyxiant by
displacing the amount of air necessary to
sustain life. This gas is extremely inert,
except when heated to very high
temperatures where it combines with
metals to form nitrides. At pressures of
400 kPa (4 atm) or higher, the gaseous
nitrogen in normal air induces a narcotic
action evidenced by decreased ability to
work, mood changes and frequently a
128
Leak Testing
mild to marked euphoria. These responses
are similar to those associated with
alcoholic intoxication.
Precautions with Nitrous
Oxide
Nitrous oxide (N2O) is used as a tracer gas
in the performance of some leak tests,
such as those using the infrared leak test
method. Nitrous oxide is a colorless,
nonflammable gas with a slightly sweetish
taste and odor. It is nontoxic and
nonirritating and must not be confused
with other nitrogen oxides that can be
harmful. Nitrous oxide is a rather weak
anesthetic and must be inhaled in high
concentrations, mixed with air or oxygen,
when regularly used as an anesthetic in
medicine and dentistry. Medical and
dental personnel who repeatedly inhale
this gas over a long period of time are
known to suffer nerve damage. When
inhaled without oxygen, nitrous oxide is a
simple asphyxiant. Inhalation of small
amounts of nitrous oxide often produces a
type of hysteria, which accounts for its
common name of laughing gas.
It is to be recognized that most other
nitrogen oxides can be harmful.
California’s Occupational Safety and
Health Administration standard for all
nitrogen oxides combined is a
concentration of 5 µL·L–1 ceiling for an
8 h occupational standard. The California
ambient air standard for nitrogen-oxygen
pollutants is 0.25 µL·L–1 for 1 h. The
Federal standards in 1978 for nitrogen
oxides, determined as time weighted
averages (TWAs), are 25 µL·L–1 or
30 mg·m–3 for nitric oxide (NO), 5 µL·L–1
or 9 mg·m–3 for nitrogen dioxide (NO2)
and 2 µL·L–1 or 5 mg·m–3 for nitric acid
(HNO3).
Precautions with Oxygen
Even though oxygen (O2) is not often
used as a tracer gas, there should be full
awareness of its potential hazards. Oxygen
is a colorless, odorless, tasteless gas and its
outstanding properties include its ability
to sustain animal life and to support
combustion. Inhalation of 100 percent
oxygen at atmospheric pressure (100 kPa
or 1 atm) will irritate the throat although
symptoms of oxygen poisoning do not
occur if the exposure is relatively short.
Long periods of exposure to higher
oxygen pressures can adversely affect
neuromuscular coordination and the
power of attention. Inhalation of oxygen
when its partial pressure exceeds 200 kPa
(2 atm) may result in the signs and
symptoms of oxygen poisoning. These
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
include tingling of fingers and toes,
acoustic hallucination, confusion, muscle
twitching (especially about the face) and
nausea. The final result of such exposure
may be convulsion, which ceases as soon
as exposure to high partial pressures of
oxygen is terminated. Note that carbon
dioxide enhances the toxicity of oxygen
and the narcotic effect of nitrogen.
Precautions against Oxygen Fires and
Explosions
Pressurized oxygen reacts violently with
oil, grease, fuel gases or metallic particles,
often producing flames or violent
explosions. The cylinders in which
gaseous oxygen is supplied are often
pressurized to 14 or 15 MPa
(2.2 × 103 lbf·in.–2 gage). Thus, oil, grease
or readily combustible materials should
never be allowed to come into contact
with interiors of oxygen cylinders, valve,
pressure regulators and fittings. These
components should never be lubricated
with oil, grease or other combustible
substances containing hydrocarbons.
Oxygen gages, regulators and fittings
should never be used for compressed air
(which may contain lubricants from air
pumps). Similarly, gages regulators and
fittings used with air or other gases should
never be used on oxygen systems, for fear
of violent explosions. It is also advisable
never to use manifolds for pressurized
oxygen systems unless these are designed
and constructed with the advice and
control of a qualified engineer. Manifolds
must comply with applicable regulations
and safety procedures. Cylinders of
oxygen should not be stored near
cylinders of acetylene or fuel gases.
Characteristics of Sulfur
Dioxide
Sulfur dioxide (SO2), through extremely
undesirable, is sometimes used in the leak
testing of welded pressure vessels. It is a
highly irritating, nonflammable, colorless
gas at room temperature and atmospheric
pressure. Liquid sulfur dioxide may cause
skin and eye burns on contact with these
tissues as a result of the freezing effect of
sulfur dioxide liquid on the skin or eyes.
Sulfur dioxide is also a highly irritating
gas in the vapor form, but is readily
detectable in concentrations of 1 to
3 µL·L–1 and so provides ample warning of
its presence.
Slight tolerance, at least up to the odor
threshold and general acclimatization are
common. Sensitization in a few
individuals, particularly young adults,
may develop following repeated exposure.
In higher concentrations, the severely
irritating effects of gaseous sulfur dioxide
make it unlikely that any person would be
able to remain in such a contaminated
atmosphere unless he or she were
unconscious or trapped. The adverse
effects of sulfur dioxide are heightened by
the presence of dust, dirt, soot or other
particulates in the air. If particulates are
high in concentration in the air, even a
little sulfur dioxide can cause illness.
Chronic exposure to sulfur dioxide
may result in fatigue, altered sense of
smell and chronic bronchitis symptoms.
Short acute exposure to sulfur dioxide gas
has severe effects. A concentration of 8 to
12 µL·L–1 causes throat irritation,
coughing, constriction of the chest, tears
and smarting of the eyes; a concentration
of 150 µL·L–1 causes extreme irritation and
can be tolerated for only a few minutes;
and a concentration of 500 µL·L–1 causes a
sense of suffocation because it is so
acutely irritating. Acute overexposure to
sulfur dioxide may result in death from
asphyxiation.
Precautions with Sulfur Dioxide
Sulfur dioxide should be handled only in
a well ventilated area, preferably using a
hood with forced ventilation. Personnel
handling sulfur dioxide should wear
chemical safety goggles or plastic face
shields (or both), approved safety shoes
and rubber gloves. Additional gas masks,
airline gas masks and self-contained
breathing apparati should be at hand for
emergencies. Instant acting safety showers
should be available in convenient
locations.
Where sulfur dioxide gas is excessive,
the worker should be supplied with a full
face piece cartridge, canister respirator or
supplied air respirator. Goggles, protective
clothing and gloves should be worn if
splashes of liquid are likely. In areas of
splash or spill, impervious clothing
should be supplied. If work clothes are
wetted by sulfur dioxide, they should be
removed promptly and the skin area
washed thoroughly.
Safety Aspects of Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
129
PART 6. Safety Precautions with Compressed
Gas Cylinders
Handling and Use of
Compressed Gas Cylinders
Most of the gas used for leak testing is
purchased in cylinders, which should be
constructed and maintained in
accordance with regulations of the
Interstate Commerce Commission. The
contents should be legibly marked on
each cylinder in large letters.
Serious accidents may result from the
misuse, abuse or mishandling of
compressed gas cylinders. Technicians
assigned to the handling of pressurized
cylinders should be carefully trained and
work only under competent supervision.
Observance of the following rules will
help control hazards in the handling of
compressed gas cylinders.
1. Accept only cylinders approved for use
in interstate commerce for
transportation of compressed gases.
2. Do not remove or change numbers or
marks stamped on cylinders.
3. Never move cylinders unless the
protective cap is in place. Because of
their shape, smooth surface and heavy
weight, cylinders are dangerous to
carry by hand and some type of
carrying device should be used when
they must be moved without the aid
of a cart. Cylinders may be tilted and
rolled on the bottom edge, but they
should never be dragged.
4. Protect cylinders from cuts or
abrasions.
5. Do not lift a compressed gas cylinder
with an electromagnet. Where
cylinders must be handled by a crane
or derrick when testing field erected
vessels, carry them in a cradle or
similar device. Take extreme care that
they are not dropped. Do not use
slings or chains.
6. Do not drop cylinders or let them
strike each other violently.
7. Do not use cylinders for rollers,
supports or any purpose other than to
contain gas.
10. When empty cylinders are to be
returned to the vendor, mark them
EMPTY or MT with chalk. Close the
valves and replace the valve protection
caps.
11. Load cylinders to be transported so as
to allow as little movement as
130
Leak Testing
possible. Secure cylinders to prevent
violent contact or upsetting.
12. Always consider cylinders as full and
handle them with corresponding care.
Accidents have resulted when
containers under partial pressure were
thought to be empty.
13. Use of safety chains to secure cylinders
during use to prevent accidental
falling is required practice by the
Occupational Safety and Health
Administration.
Precautions for Storage of
Compressed Gas Cylinders
Store compressed gas cylinders with
protective caps properly installed in safe,
dry and well ventilated places prepared
and reserved for this specific purpose.
Cylinders should be stored on a level,
fireproof floor and should be chained in
place or provided with barriers to prevent
them from falling over. Flammable
substances such as oil and volatile liquids
should not be stored in the same area as
pressurized gas cylinders. Cylinders
should not be stored near arc welding
areas, elevators, gangways, stair wells or
other places where they could be knocked
over, arc gouged or damaged. Cylinders
are not designed for temperatures in
excess of 55 °C (130 °F). Accordingly, they
should not be stored near sources of heat
such as radiators or furnaces, nor near
highly flammable substances like gasoline.
Cylinder storage should be planned so
that cylinders will be used in the order in
which they are received from the supplier.
Empty and full cylinders should be stored
separately, with empty cylinders being
plainly identified as such to avoid
confusion. Group together cylinders that
have held the same contents.
Precautions in Indoor
Storage of Oxygen and
Fuel Gas Cylinders
Cylinders of oxygen must not be stored
indoors close to cylinders containing
flammable gases. Unless they are stored
apart, oxygen cylinders and flammable
gas cylinders must be separated by a fire
resistive partition. A direct flame or
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
electric arc should never be permitted to
contact any part of a compressed gas
cylinder.
Acetylene and liquefied fuel gas
cylinders should be stored with the valve
end up. The total capacity of acetylene
cylinders stored inside a building should
be limited to 60 m3 (2000 ft3) of gas,
exclusive of cylinders in use or connected
for use. Quantities exceeding this total
must be stored in a special room, located
in a separate building or outdoors and
built in accordance with the specifications
of NFPA 51, Standard for the Design and
Installation of Oxygen-Fuel Gas Systems for
Welding, Cutting, and Allied Processes.13
Storage rooms for cylinders containing
flammable gases should be well ventilated
to prevent the accumulation of explosive
concentrations of gas. No source of
ignition will be permitted; smoking must
be prohibited. Wiring should be in
conduit. Electric lights should be in fixed
positions and enclosed in glass or other
transparent material and equipped with
guards to prevent breakage. (Note that
glass enclosures, electrical conduit and
conventional switch and receptacle boxes
used in electrical wiring systems do not
prevent entry of gases into their
enclosures.) Therefore, electrical switches,
which are subject to sparking or arcing
during operation, should be located
outside the room in which flammable
gases are stored.
Precautions in Outdoor
Storage of Gas Cylinders
One common type of storage house
consists of a shed roof with side walls
extending about halfway down from the
roof and a dividing wall between
cylinders of one kind of gas and those for
another gas. To prevent rusting, cylinders
stored in the open should be protected
from contact with the ground and against
extremes of weather, accumulations of ice
and snow in winter and continuous direct
rays of the sun in summer.
Safe Procedures for Using
Cylinders of Compressed
Gases
Safe procedures for compressed gas
cylinders include the following.
1. Use cylinders in the upright position
and secure them to prevent them from
being accidentally knocked over.
2. Unless the cylinder valve is protected
by a recess in the head, keep the metal
cap in place to protect the valve when
the cylinder is not connected for use.
A blow on an unprotected valve might
cause gas under high pressure to
escape.
3. Make sure the threads on a regulator
or union correspond to those on the
cylinder valve outlet. Do not force
connections that do not fit.
4. Open cylinder valves slowly. A
cylinder not provided with a
handwheel valve should be opened
with a spindle key, a special wrench or
other tool provided or approved by
the gas supplier.
5. Do not use a cylinder of compressed
gas without a pressure reducing
regulator attached to the cylinder
valve, except where cylinders are
attached to a manifold, in which case
the regulator should be attached to
the manifold header.
6. Before making connection to a
cylinder valve outlet, except that of a
hydrogen cylinder, crack the valve for
an instant to clear the opening of
particles of dust and dirt. Always point
the valve and opening away from the
body and not toward anyone else.
Operators should wear safety glasses.
7. Use regulators and pressure gages only
with gases for which they are designed
and intended. Do not attempt to
repair or alter cylinders, valves,
regulators or attachments. This work
should be done only by the
manufacturer.
8. Unless the cylinder valve has first been
closed tightly, do not attempt to stop
a leak between the cylinder and the
regulator by tightening the union nut.
9. Combustible gas cylinders in which
leaks occur should be taken out of use
immediately and handled as follows:
(a) Close the valve and take the
cylinder outdoors well away from any
source of ignition. Properly tag the
cylinder and notify the supplier. A
regulator attached to the valve may be
used temporarily to stop a leak
through the valve seat. (b) If the leak
occurs at a fuse plug or other safety
device, take the cylinder outdoors well
away from any source of ignition,
open the cylinder valve slightly and
permit the gas to escape slowly. Tag
the cylinder plainly. Post warnings
against approaching with lighted
cigarettes or other sources of ignition,
promptly notify the supplier and
follow its instructions for returning
the cylinder.
10. Do not permit heavy objects, sparks,
molten metal, electric currents,
excessive heat or flames to come in
contact with cylinders or attachments.
11. Never use oil or grease as a lubricant
for valves or attachments of oxygen
cylinders. Keep oxygen cylinders and
fittings away from oil and grease and
do not handle such cylinders or
Safety Aspects of Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
131
apparatus with oily hands, gloves or
clothing. Signs should be posted
where oxygen is stored, prohibiting
oil, grease or other lubricants on
oxygen equipment.
12. Never use oxygen as a substitute for
compressed air in pneumatic tools or
to start internal combustion engines
or for pressurizing a system for testing
or for dust removal. Use it only for the
purpose for which it is intended.
13. Never bring gas cylinders into vessels
or unventilated rooms.
14. Do not fill cylinders except with the
consent of the owner and then only in
accordance with regulations. Do not
attempt to mix gases in a compressed
gas cylinder or to use it for purposes
other than those intended by the
supplier.
15. Secure all gages and hoses with proper
size wrenches, not slip jaw pliers.
16. Do not overtighten or strip threads on
cylinder attachments.
different colors. Cylinder valve outlet
threads have been standardized for most
industrial and medical gases by the
American National Standards Institute,
recommending different combinations of
right hand and left hand threads, internal
and external threads and different
diameters to guard against wrong
connections. Standards are being rapidly
adopted whenever gas manufacturers and
industrial users reach agreement to
change both valve outlets and regulator
connections. Adaptors are used in the
interim until the changes are completed.
The regulator is a delicate apparatus
and should always be handled carefully. It
should not be forced, dropped or
pounded. Regulators should be sent to the
manufacturer for repairs and testing by
skilled personnel.
Safety Precautions with
Valves or Regulators on
Gas Cylinders
Leaky or creeping regulators are a source
of danger and should be withdrawn from
service at once for repairs. If a regulator
shows a continuous creep, indicated on
the low pressure (delivery) gage by a
steady buildup of pressure when the
outlet valves are closed, the cylinder valve
should be closed and the regulator
removed for repairs. If the regulator
pressure gages have been strained so that
the pointers do not register properly, the
regulator must be repaired at once. When
regulators are connected but are not in
use, the pressure adjusting device should
be released. Cylinder valves should never
be opened until the regulator is drained of
gas and the pressure adjusting device on
the regulator is fully released.
Regulators or reducing valves must be
used on gas cylinders to maintain a
uniform gas supply. Technicians should
stand to one side and away from regulator
gage faces when opening cylinder valves.
Always wear safety glasses to protect eyes
from ejected particles.
Only regulators listed or approved by
agencies such as Underwriters’
Laboratories, Incorporated, should be used
on cylinders of compressed gas. Each
regulator should be equipped with both a
high pressure (contents) gage and a low
pressure (working) gage.
Safety Procedures for
Leaky or Anomalous
Regulators
Safety Precautions with
Oxygen Pressure
Regulators
High pressure oxygen gages should have
safety vent covers to protect the operator
from broken glass in case of an internal
explosion. Each oxygen gage should be
marked OXYGEN—USE NO OIL.
Serious, even fatal accidents, have
resulted when oxygen regulators have
been attached to cylinders containing
combustible gas or vice versa. To guard
against this hazard, it has been customary
to make connections for oxygen
regulators with right hand threads and
those for combustible gases such as
acetylene with left hand threads, to mark
the gas service on the regulator case, and
to paint the two types of regulators
132
Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
PART 7. Safety Precautions in Pressure and
Vacuum Leak Testing
Safety Considerations in
Leak Testing
When a pressure or a vacuum vessel is
fabricated, some means of testing must be
used to predict safe performance of the
vessel. It is sometimes necessary to exceed
the designed operating conditions during
initial pressure testing. This requires many
safety considerations to ensure proper
protection of personnel. (Hazards related
to toxic or flammable solvent vapors and
tracer gases in leak testing should also be
given careful consideration.)
Explosion and Implosion
Hazards in Pressure and
Vacuum Leak Testing
Pressurized vessels can fail by explosion
because of the energy stored in air or
nonflammable gases used to pressurize
systems during leak testing. In systems
that are evacuated during leak testing,
implosion (violent collapse) failures can
result from external (atmospheric)
pressures applied to structures not
designed for such loading. Where
flammable tracer gases are used in leak
testing in the presence of air or oxygen,
violent combustion or explosive chemical
reactions can occur. These hazards must
be foreseen and carefully controlled to
ensure safety during leak testing.
Precautions in Selecting
Sites for Leak Testing
Major factors determining the size, shape
and type of buildings and structures to be
used for leak testing of components need
to be investigated. Catastrophes resulting
in large loss of life and heavy property
damage often are due to inadequate
planning stage considerations. High
hazard leak testing operations should be
located in small isolated buildings of
limited occupancy. Buildings can be
designed so that internal explosions will
produce minimum damage and minimum
broken glass. Lower hazard operations can
justify large units.
Pressure Vessel Code
Requirements for Safety
Procedures
The degree of safety precautions necessary
during leak testing varies greatly with the
type of system being tested. In the case of
hydrostatic and pneumatic tests of
pressure vessels, the ASME Boiler and
Pressure Vessel Code outlines the minimum
safety procedures to be followed during
pressure testing. The ASME Boiler and
Pressure Vessel Code and other applicable
specifications should be followed with
care to ensure safety in all operations to
which they apply. However, often it is the
rather subtle hazard that may be
disastrous. Potential hazards should be
taken into account both when preparing
for or performing leak testing. These
include tracer gas safety aspects such as
flammability, asphyxiation or specific
physiological effects as well as the
possibility of pressure vessel explosions.
Protecting Test Personnel
during Pressure Testing
Greater respect for high pressure testing
has led to increased emphasis on safety,
with the result that overall safety
experience has been very good. This
respect is well justified when one realizes
that a valve stem operating at 200 MPa
(3 × 104 lbf·in.–2) that fails and is blown
out is propelled under conditions similar
to those of a bullet fired from a high
powered rifle. The energy released from a
completely liquid system should not be
underestimated either. Compressed liquid,
although smaller volumetrically than
compressed gas, is very much to be
reckoned with in considering potential
forces to be handled when pressure is
released. For example, a gasket 0.4 mm
(0.016 in.) thick, blown between split
flanges under a pressure of more than
10 MPa (more than 2 × 103 lbf·in.–2), will
release a thin sheet of water like a knife
edge that could cause injury, eye damage
and loss of sight.
Successful personnel protection during
pressure testing involves not only
mechanical devices to guard against
Safety Aspects of Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
133
injury should failure occur, but thorough
training of people, establishment and
enforcement of rigid safety rules and
necessary disciplinary action when
justified. Without the proper attitude and
respect for what is being handled, trouble
is sure to occur.
Safety with Scaffolds
A scaffold is an elevated working
platform, usually temporary, for
supporting both men and materials. For
safety’s safe, scaffolds should be designed
to support at least four times the
anticipated weight of men and materials
to be placed on them and all elevated
working platform areas should be guarded
(as by railings) on all exposed sides.
Working scaffolds should not be used as a
platform for jacking or leverage purposes
without proper allowance for the added
loads and stresses.
Barricades, Protective
Walls and Distance for
Safety during Leak Testing
Based on safety experience accumulated
during laboratory operations and on
sound design principles, a custom vessel
can be built with reasonable assurance
that it may be leak tested or pressure
tested safely. While complete isolation
usually is not required, certain pieces of
equipment may need barricade
protection. Access to the test area during
testing should be restricted to minimize
exposure of personnel to hazards.
Remote control and observation may
be used where possible during leak
testing. Periscope techniques, shatterproof
glass windows and industrial television
offer opportunities to check on operating
equipment without exposure. Instrument
data can be transmitted electrically or by
low pressure pneumatic systems to a
separate control room. Valves that are not
controlled automatically can be operated
by rods or shafts extended through a
barricade gage board combination with
proper seals.
Pressure Vessel Design and
Causes of Failures
Fired and unfired pressure vessels of many
types are in common use in industrial,
commercial and public buildings for space
and process heating and heat exchange;
for processing food, chemicals, petroleum
and other industrial products; and for
processes involving nuclear energy. These
vessels hold gases, vapors, liquids and
134
Leak Testing
solids at various temperatures and at
various pressures, ranging from absolute
pressures of nanopascals or lower
pressures to tens of megapascals (10–9 to
107 Pa).
Some common causes of failure in
pressure vessels are the following:
(1) errors in design, construction and
nondestructive testing; (2) improper
education of testing personnel;
(3) mechanical breakdown such as failure,
blocking or lack of safety devices; (4) poor
visual inspection before pressurization;
(5) improper test procedure; (6) improper
application of test equipment; (7) blocked
or dysfunctional gages; (8) test flanges or
valves of wrong material; (9) improperly
designed test flanges; and (10) test
pressure too high. These causes of
potential failures should be anticipated
and avoided insofar as possible. Before a
pressure vessel is tested, three questions
should be answered about its design.
1. Can the filled vessel carry the weight
of its contents in addition to the
internal pressure without undue
strain?
2. Can the support structure and
building floor carry the weight of the
filled vessel?
3. Can the vessel withstand any vacuum
and not collapse under external
atmospheric pressure that may be
created either accidentally or
intentionally?
It is imperative that any safety enclosures
be designed to withstand the worst
possible conditions of failure; otherwise,
protective walls may break and lethal
fragments of metal or concrete can be
blown outward. Vented roofs or pressure
testing below ground level should be
considered when pressure testing with
compressed air or gases.
Precautions for Protection
against Equipment Failure
from Overpressure
Safety precautions to protect personnel
and equipment from failures during
pressure testing include the following.
1. Ensure that the test equipment and
vessel under test are properly designed
and constructed in the first place.
2. Before pressure testing, ensure that
equipment is properly assembled to
avoid overstressing. This includes
proper bracing and shoring under
pressure vessels to support critical
points. Otherwise it is possible that
failure may actually be started before
or while equipment is being set up for
the test.
3. Be sure that careful visual and other
inspections are done during
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
construction and testing to guarantee
compliance with design, proper
manufacturing procedures, material
choices and workmanship standards.
4. Watch out for areas where stress
concentrations or nonuniform loading
in enclosures, pump and compressor
cylinders, valves etc. may cause
sudden or gradual failures.
5. Install preliminary warning devices
that alert the leak testing technician
when test pressures are increasing too
rapidly or when pressurization is
approaching an excessive level. These
devices can call attention to an
abnormal situation before a pressure
relieving device is set off. Prompt
correction of trends toward excessive
pressure can often forestall the
actuation of emergency pressure relief
valves. This is most valuable in
extreme pressure work.
6. Check temperature of test water or
other test medium for compliance to
test procedure.
7. Assure the availability on test site of
approved written test and safety
procedures for all test personnel.
Pressure Relieving Devices in Pressure
Leak Testing
Spring loaded relief valves are used up to
100 MPa (1.5 × 104 lbf·in.–2) as pressure
relieving devices. They are quite reliable
for nonpulsating operations at 15 to
20 percent above working pressure, but
cannot be completely relied on to reseat
without leakage. Shear rupture disks,
made of bronze, stainless steel or other
metals, depending on service conditions,
are suitable for nonpulsating operations at
test pressures up to 20 to 30 percent
above working pressure. Formed heads
failing in tension have been applied to
appreciable pressures but do not possess
the accuracy required at higher pressure
up to 70 MPa (1 × 104 lbf·in.–2).
Sometimes relief valves and rupture
disks are used in parallel. In this case, the
relief valve is set to open at a lower
pressure. This warns test technicians that
prompt corrections may be necessary to
avoid rupture disk failure, with resulting
lost time. Rupture disks and relief valves
are also used in series, with one or the
other in the upstream position. In this
series case, unless a small vent hole is
used between the two to prevent seepage,
the back pressure caused by seepage can
force the failure pressure on the upstream
unit to rise to a dangerous value equal to
the relieving pressure.
Hydraulically loaded plugs using O-ring
seals are dependable and will relieve at
test pressures closer to the working
pressure than other devices. O-ring seals
are typically flexible ring shaped inserts
placed in circular grooves and compressed
to form tight seals between mating parts
of pressure or vacuum systems. They can
have any cross sectional area required for
protection and can be designed for any
relieving pressure. The O-ring seals should
be made of material that will not fail or
deteriorate from the test medium used.
Pressure Gage Calibration
and Safety Applications
One of the best means of protection from
overpressure is to use an accurate gage. To
ensure accuracy of a pressure gage, it must
be periodically checked against some
known standard pressure. Dead weight
testers are used for calibration and
checking of the elastic gages for pressures
exceeding approximately 100 kPa
(15 lbf·in.–2) and extending to 70 MPa
(105 lbf·in.–2) or even higher. Dead weight
gages are used for the precise
determination of essentially constant
pressures maintained in a vessel by some
pressure generating mechanism. The dead
weight tester and the pressure gages
should both be calibrated over their full
scale. Pressure gages should be calibrated
both before and after testing on critical
high pressure tests. Gage calibration
should follow approved written
procedures.
Care, Handling and
Storage of Pressure Gages
Handling and storage should be done
with the knowledge that a gage suitable
for accurate pressure measurement is as
delicate as a watch. Its removal and
replacement for calibration purposes (and,
of course, its installation and use) should
be entrusted only to persons who can be
depended on to avoid dropping or jarring
the gage or subjecting it to rough
treatment. The gage should always be
attached by using a wrench on the flats
provided on the connection. A gage must
never be screwed or unscrewed by using
the gage casing. If the gage has the proper
tolerances and is handled correctly, the
gage corrections determined before and
after the pressure test for each test for
which the gage was used should agree
within specified calibration accuracies. If
the gage is not handled properly, there is
a chance that the calibration and
corrections determined before and after
the test will differ appreciably. In such
events there is no sure way to know
which correction to use and the result of
the test will be in doubt.
Any accident to a gage requires that
the gage be given a complete calibration
Safety Aspects of Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
135
and correction test before further use. This
would apply also if a gage shows obvious
evidence of prior damage. In the event
that a gage is mishandled by dropping it,
exceeding its pressure range or exposing it
to vacuum (unless it is an absolute
pressure gage), the gage must be repaired
and recalibrated by its manufacturer, a
qualified laboratory or equipment
manufacturer with proper calibration
facilities.
Hazards of Pressurized
Test Systems
The necessary safety precautions vary
greatly with the type of system being leak
tested. Some general types are listed below
in ascending order of the potential danger
involved.
1. With small hydraulic systems of
moderate pressure, the major hazard is
from a jet of the liquid either from a
leak or failure. Occasionally, the
necessity to include a brittle material
such as a sight glass or glass flow
meter in the system adds the hazard of
flying particles.
2. Low pressure systems involving
nonreactive gases or liquids above
their boiling point involve little
hazard if correctly handled. However,
it is important to have the proper
relief valves, rupture disks and
pressure regulators to maintain safety
in low pressure systems. The hazard of
low pressure systems can be higher if
large volumes of gases are involved.
3. Systems involving flammable gases or
liquids (such as kerosene) as the
pressure testing fluid involve major
hazards, including those of fires or
explosions resulting from leakage or
failure of some component.
4. The hazards of high pressure hydraulic
and inert gas systems increase with the
increase in pressure, the
compressibility of the testing media
and the volume of the system. There is
an increasing probability that
equipment in the higher pressure
ranges will not permanently resist the
effect of pressure.
Explosion of Systems or
Vessels Pressurized for
Leak Testing
If a system to be leak tested is pressurized
with tracer gas or gas mixtures, rupture of
its containment walls or pressure
boundaries could produce considerable
damage. If the system being pressurized is
small, it might seem as if few precautions
136
Leak Testing
would be necessary during pressurizing.
However, the damage from rupture of a
gas filled volume results from the total
amount of gas it contains. Therefore,
either a small system under high pressure
can be as dangerous a large system under
lower pressure. The energy stored in a
pressurized gas volume is equal to the
product of its pressure and its volume.
The pressure in pascal or newton per
square meter multiplied by the volume in
cubic meter results in energy in joule,
(N·m–2) × m3 = Nm = J. By comparison,
1 kg of gasoline contains about 44 MJ,
enough to blow up a tank.
When pressurizing a system, a pressure
regulator fitted with a safety overpressure
release device should be installed so that a
pressure in excess of the design pressure
can never be applied to a vessel or system
under test.
Rupture Hazards in
Pressure Testing
Although the prevention of clogged leaks
dictates that leak testing with gaseous
tracers should be done before contact of
the system with liquid, the need for safety
might overrule this procedure.
Pressurizing a system with a liquid does
not create the explosion hazard involved
with gases under high pressure. Therefore,
safety requirements may dictate pressure
testing of a system with a liquid before
gases are introduced for leak testing. An
alternate preliminary leak test might be a
low pressure, high sensitivity mass
spectrometer leak test using helium as the
tracer gas.
The amount of energy stored in a tank
pressurized with gas is a function of the
quantity and type of gas contained in the
vessel. Because of this, a high volume, low
pressure vessel can contain the same
stored energy as a low volume, high
pressure vessel. Therefore, each presents a
hazard of similar magnitude.
The exact rupture hazard involved in
pressure testing is difficult to define,
although the structural burst limit is a
reasonably predictable design factor. Any
damage incurred during fabrication,
erection, testing or service by a vessel
under pressure, such as weld undercut or
a deep nick or gouge, may cause explosive
failure if the damage is severe. Small flaws
can be progressive depending on metal
strain and the type of load. Surface stress
concentrations caused by vessel damage
may not result in immediate failure, but
may progress and cause failure later.
When a skin puncture takes place, it
results in a tearing action that tends to
enlarge the hole. An inspection of a failed
vessel will show tears extending across the
entire face of the skin.
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
LT.04 LAYOUT 11/8/04 2:15 PM Page 137
Effects of Leak Size and Shape of
the Opening on Failure
Mechanisms
A smaller leak will dissipate the same
energy as a larger leak, but over a long
period without the explosive effect. The
critical leak size is related to the tensile
strength of the enclosing skin. The critical
point of explosive pressure release is
reached when the force of the gas
escaping from the hole exceeds the force
that can be withstood by the skin.
Another factor influencing failures is
the shape of the leak opening. An
irregularly shaped opening, with many
microscopic irregularities, each providing
a stress point, offer an ideal starting point
for a tearing and shredding action. As the
pressure in a tank increases, a critical
pressure is reached where the stress exerted
by the confined gas exceeds the strength of
the metal surrounding the failure. This
causes an explosive disintegration.
Variation of the critical point of explosive
pressure release can occur with conditions
of service, vessel shape, size and wall
thickness and material, fabrication
methods and type of failure.
Energy Contained in Pressurized
Vessels
The work done to compress gas in a vessel
is stored in that gas. It is normally
returned by propelling the gas to places
where it is needed. However, a rupture in
the tank may suddenly release all the
energy at once as an explosion. An
explosion is so fearsome because of the
short duration of the energy release that
can be calculated by means of equations
for isentropic processes:
(3)
PV
=
k
constant
and for work 1W2:
(4)
1W 2
=
pressure P2 equals 100 kPa (one
atmosphere), the energy released can be
computed as follows. Because PVk is a
constant,
1
(5)
 P k
= V1  1 
 P2 
V2
1
Equations 3 and 4 take into account
any sudden temperature change during
the explosion. In this equation, k = 1.4, a
constant. The work done (energy released)
is that resulting from a change from
conditions identified by subscript 1 to
those identified by subscript 2. As an
example, compute the energy released
when a small pressurized tank is ruptured
and compressed gas escapes to
atmospheric pressure and temperature. If
the internal pressure within the tank is
P1 = 15 MPa (150 atm), the volume of the
tank is V1 equals 0.04 m3, and the
compressed gas escaping to atmospheric
=
1.43 m 3
Substituting in the work equation:
1W2
=
(100)(1.43) − (15 000)(0.04)
=
1.142 MJ
1 − 1.4
This energy (somewhat more than
1 MJ) could be evaluated in comparison
with the 44 MJ of energy available by
combustion of 1 kg of gasoline, of 38 MJ
from 1 m3 of natural gas or of 32 MJ from
1 kg of coal.
Evaluating Hazards of
Explosive Pressure Release
The critical point of explosive pressure
release is a very important factor in
determining the hazard magnitude of
high pressure leak tests. However,
calculation of available stored gas energy
is necessary for a thorough analysis of the
potential hazard. This calculation includes
two important considerations: (1) the
amount of energy stored in the
compressed gas and (2) the rapidity with
which this energy is released.
The amount of energy stored in a
noncombustible compressed gas can be
approximated by Eq. 6:
P2V 2 − P1V1
1 − k
=
 15 000  1.4
0.04 

 100 
(6)
E
=
K −1



 P  K
 2 
− 1

1 − K  P1 




P1 V1
where K is the ratio of specific heat Cp at a
constant pressure to that of a constant
volume Cv, where P1 is initial absolute
pressure, V1 is initial volume and P2 is
final pressure (100 kPa or 1 atm). This
equation is based on the ideal gas law and
isentropic expansion. At high pressure
(e.g., above 20 MPa) where the deviation
from an ideal gas may be appreciable, the
equation is still valid provided one divides
the right hand side by 2, the so-called
compressibility factor found in gas
handbooks.
Safety Aspects of Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
137
Applying Eq. 3 to helium, for example,
one obtains a stored energy of 1.1, 12 and
106 MJ·m–3 for initial pressures of 1, 10
and 100 MPa, respectively. The stored
energy can be converted to
2,4,6-trinitrotoluene (TNT) equivalents by
using the conversion factor of
2.38 × 10–10 tons of TNT per joule. Also
evidenced in Eq. 3 is the fact that a high
volume, low pressure vessel can contain
the same stored energy as a low volume
high pressure vessel. Therefore, it can
present a hazard of similar magnitude.
Of critical importance is the rapidity
with which the energy release occurs. For
the purpose of hazard definition, the
extreme case of total and instant removal
of gas confinement is used. The sudden
release of energy is transmitted through
the air in the form of a shock wave,
generated by the sudden displacement of
air surrounding the vessel. The shock
wave carries with it measurable
overpressure, varying with the intensity of
the initial displacement. This shock wave,
however, is never greater than the
pressure that caused the displacement.
The shock wave diminishes as a factor of
distance.
Human Injury from Shock
Wave Overpressures
It has been established that no damage
will occur to a human body when it is
subjected to shock overpressure of not
more than 17 kPa (2.5 lbf·in.–2). Body
displacement can occur with shock wave
overpressures of 20 to 35 kPa (3 to
5 lbf·in.–2). However, a human body can
be subjected to shock overpressures as
high as 35 kPa (5 lbf·in.–2) without injury
to the internal organs. Above 35 kPa
(5 lbf·in.–2), eardrum rupture can occur.
Permanent lung damage will be
experienced with shock wave
overpressures of 100 kPa (15 lbf·in.–2 or
1 atm). Fatalities will occur with
increasing probability with shock wave
overpressures above 250 kPa (35 lbf·in.–2).
The distance from the source of a
shock wave at which personnel will be
subjected to 35 kPa (5 lbf·in.–2)
overpressure is selected as the minimum
safe distance. Because injury can occur
from body displacement against the
ground or nearby structures, personnel
must be protected from direct exposure to
a 35 kPa (5 lbf·in.–2) shock. No one except
the minimum crew necessary to conduct
leak tests should be allowed inside the
area when the vessels are being
pressurized.
138
Leak Testing
Hazards of Vacuum Testing
Evacuated systems, while not generally
considered hazardous, involve the dangers
of implosion or the possibility of
personnel entering a vessel which, even
though it has been vented to the
atmosphere, does not contain enough
breathable air to sustain life. Most
vacuum testing involves gases such as
helium, nitrogen and hydrogen, which
will not support life. The same general
precautions of handling pumping
equipment, compressed gases, sight
glasses etc. apply to vacuum testing as
well as pressure testing.
Hazard of Implosion of
Systems of Vessels
Evacuated for Leak Testing
Implosion is the collapse of a pressure
boundary or the walls of a containment
vessel or structure when evacuated and
subjected to atmospheric or higher
external pressures. Many vessels and
chambers are made for use under vacuum
to simulate high altitude or outer space
conditions where the maximum pressure
differential that will ever be applied across
their boundaries is 100 kPa (1 atm) of
external pressure. Systems fabricated of
thin wall materials, glass or foils cannot
withstand high external or internal
pressures.
For example, although they are not
internally pressurized, glass bell jars that
are evacuated can become a dangerous
source of flying glass as a result of
implosions. Pieces of flying glass,
propelled by a pressure difference of about
100 kPa (1 atm), will travel great distances
unless they should happen to collide with
a safety shield or glass pieces coming from
the opposite direction. The hazard of
personnel injury by flying glass becomes
particularly serious when the capacity of
the glass vessel exceeds about 30 L (1 ft3).
For this reason, all evacuated bell jars
should be enclosed in some type of safety
shield.
Safety shields should be used on small
thin wall vessels and glass bell jars under
all vacuum conditions if an implosion
hazard exists. The pressure differential
between atmospheric pressure (101 kPa or
1.01 atm) and an absolute pressure of a
typical vacuum (100 Pa or 0.001 atm) is
essentially equal to atmospheric pressure
(100 kPa or 1 atm). Any additional
increase in pressure differential is
negligible as the contained vacuum is
further evacuated from 100 Pa to 1 Pa.
Most of the atmospheric pressure is thus
exerted on the bell jar or thin wall system
when rough evacuation takes place. The
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
increase in pressure difference resulting
from further pumping to obtain a high
vacuum is very small. Thus, it is a mistake
not to use bell jar safety shields for any
but the most moderate vacuum.
Vacuum Vessel Design
Vacuum vessel design may be divided
roughly into two parts: (1) physical
design, which is chiefly concerned with
design for strength and satisfactory
mechanical operations and (2) functional
design, which is in the realm of vacuum
engineering. Unless a thorough
understanding of all the vacuum process
variables is obtained, the finest
mechanical design will not ensure
satisfactory results when the equipment is
placed in operation. The final design of a
vessel, as in all engineering work,
represents a number of compromises
between conflicting conditions. The
designer must consider all factors
involved, both physical and functional,
and then endeavor to reach the optimum
solution. Where vacuum vessels do not
come under ASME Boiler and Pressure Vessel
Code requirements, it is recommended
that the ASME Boiler and Pressure Vessel
Code be used whenever applicable.14
Pressure Proof Testing of
Systems before Leak
Testing
Before undertaking leakage measurements,
large systems may require proof testing to
determine their capability to withstand
leak test pressurization. For example, the
ASME Boiler and Pressure Vessel Code
(Section I, “Power Boilers”; Section III,
“Nuclear Vessels”; and Section VIII,
“Unfired Pressure Vessels”)14 specifies that
all vessels should be hydrostatic proof
tested to 1.5 times the maximum
allowable working pressure.
The alternative to hydrostatic proof
testing with water is to perform a
pneumatic proof test to 1.25 times the
maximum allowable working pressure.
The pneumatic proof test may be
performed by pressurizing with gas to a
high pressure while all personnel are
removed from the test area. The
disadvantage of the proof test made with
gas or air pressure is that if the system
bursts during testing, considerable damage
can result.
The alternative to proof testing with
pressurized gas is to make a hydrostatic
pressure proof test in which the system is
pressurized with water.) Because water is
relatively incompressible under pressure
(as compared with gases), the energy
released when a system bursts under water
pressure is far less than when the system
bursts under an equal gas pressure. On the
other hand, if hydrostatic testing is
performed before leak testing with
gaseous tracers, any small leaks in the test
system will become clogged with water.
Therefore, if at all possible, hydrostatic
testing should not be performed on test
vessels or systems where the allowable
leakage rate is less than 10–7 Pa·m3·s–1
(10–6 std cm3·s–1).
Codes and Requirements
for Testing of Pressure and
Vacuum Vessels
The most valuable source of information
for the guidance of the engineer in the
matter of physical design is the ASME
Code, which is issued by the American
Society of Mechanical Engineers and
governs the design of unfired pressure
vessels.14 This ASME Code is the result of
the contributions of many authorities
representing designers, builders and users
of vessels. The ASME Code rules and
procedures have safe operation as their
fundamental objective.
The mandatory pressure or vacuum
vessel requirements of the states,
municipalities and insurance companies
involved should be studied, as it may
become necessary to have the vessel ASME
Code-stamped. Under these conditions,
the ASME Code must be adhered to, Code
calculations and design submitted to the
proper authorities for approval and the
vessel fabricated by those companies
holding an ASME Code Certification of
Authorization for the manufacturing
involved. Chemical analysis and
mechanical properties of all material that
is under ASME Code rules are required and
vessel manufacturers must have verified
material certifications from the supplier.
The ASME Code vessel or component
inspection and stamping verification must
be done by an authorized inspector
holding a valid and current National
Board commission (from the National
Board of Boiler and Pressure Vessel
Inspectors, Columbus, Ohio) in the area
involved and who is employed by an
authorized inspection agency. In addition,
the ASME Code pressure vessels must be
fabricated and manufactured under a
controlled manufacturing system and
quality assurance program as outlined in
the manufacturer’s quality assurance
manual.
Safety Aspects of Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
139
PART 8. Preparation of Pressurized Systems for
Safe Leak Testing
140
Methods for Leak Testing
of Pressurized Systems
(without Tracer Gases)
Personnel for Pressure
Testing and Leak Testing
of Pressurized Systems
Pressure vessels and pressurized systems
and components are designed to contain
fluids at atmospheric or higher than
atmospheric pressures. Pressure systems
are commonly subjected to hydrostatic,
hydropneumatic or pneumatic pressure
proof tests during their manufacture,
erection or periodic inservice
maintenance inspections. Proof tests are
made with pressurized liquids, with
liquids and gases or with gases under
pressures adequate to stress the
containment structures to ensure their
integrity. These tests often provide
evidence of locations of leaks or indicate
the presence of leakage by changes in
pressure or fluid flow rates. Similar tests
are also made on joints or sections of
transmission line pipe following welding
of longitudinal seams in pipe mills and on
completed sections of pipelines following
girth welding. Proof testing by
pressurizing is used to ensure structural
integrity and may indicate leak locations
or leak tightness.
The most sensitive leakage rate testing
is done by pressurizing the pressure
vessels, components or systems with gases
(or gaseous mixtures containing tracer
gases) to establish a pressure differential
across the containment boundary. The
rate of leakage can often be increased by
pressuring up, a technique in which
internal pressure is raised to increase the
rate of flow of gas through leaks and thus
permit faster or more sensitive leak
testing. The presence of leakage can then
be detected (1) by measurement of
pressure changes within the pressurized
system or in an enclosure containing the
pressurized components under test, (2) by
input flow rates required to maintain
pressure at constant levels or (3) by
sensitive detection of specific tracer gases
passing through the leaks.
The best equipment that can be devised
and assembled for pressure tests and leak
testing of pressure vessels and systems is
useless without properly trained and
competent leak testing personnel.
Training, although extremely necessary,
cannot take the place of intelligence and
clear thinking that is often referred to as
horse sense, ingenuity, resourcefulness,
imagination or innate ability. In addition,
special training and caution are essential
to prevent accidents or possible disastrous
pressure vessel explosions, to avoid
exposure of personnel to toxic tracer gases
and to avoid asphyxiation where
atmospheric oxygen has been displaced
by accumulations of tracer gases and
mixtures that do not support life. Where
flammable or toxic pressurizing gases or
liquids are used, full precautions must be
taken to prevent fires, explosions or
contamination of the atmosphere with
toxic gases or gases that are flammable in
air. Leak testing of pressurized systems
requires that test personnel be trained, be
intelligent and have considerable
experience in operations performed under
adverse conditions and with temporary
equipment arrangements used only
during leak testing.
Leak Testing
Development of
Techniques for Testing of
Pressure Vessels
Until recent years, leak testing of most
pressure vessels was performed in a
relatively crude manner. Hydrostatic and
pneumatic pressure tests were performed
primarily to ensure the structural integrity
of pressure vessels. Many pressure vessels
are fabricated in accordance with the
recommendations of the ASME Boiler and
Pressure Vessel Code.14 This code was
prepared by the Boiler and Pressure Vessel
Committee, established in 1911 by the
American Society of Mechanical Engineers
(ASME). The purpose of the Committee is
to formulate standard rules for the
construction of steam boilers and other
pressure vessels. The Committee
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
establishes safety rules governing the
design, fabrication and inspection during
construction of boilers and unfired
pressure vessels and interprets these rules
when questions arise regarding their
intent. The ASME Boiler and Pressure Vessel
Code provides a Standard Recommended
Guide for the Selection of a Leak Testing
Method (SE 432).15
Leakage has become a serious concern
in the fabrication of nuclear reactors and
components, as well as for vessels to
contain lethal substances. Leak testing is
also required on vessels used in the
processing of materials that are affected
by the presence of contaminants that
react with the product they contain.
Similar guides have been developed for
inspection of pressure equipment in other
industries. For example, the American
Petroleum Institute (API) provides
quidelines and recommended practices
with information on pressure vessels and
components of chemical plants and
petroleum refineries. Inspectors are
required to have complete knowledge of
the requirements and recommended
practices applicable in the specific
industry in which the pressure vessels will
be used. These inspectors often have
responsibility for both leak testing and
nondestructive testing of new
construction and of plant facilities in use
or during maintenance shutdown periods.
Mechanisms of Material
Failures at High Pressure
Many people do not realize the hazards
associated with hydrostatic testing in the
higher pressure ranges. Materials that
ordinarily are ductile can fail in a brittle
manner at low temperatures. Small defects
inherent in the grain structure, poor
quality workmanship in fabrication or
faulty design may, when the material is
stressed, start a local crack that can no
longer be arrested by the ductility and
toughness characteristics of the material.
Brittle fracture usually occurs at high
stress levels and is more likely to occur in
thick plate than in thin plate. This
thickness effect is due in part to the
increased restraint to plastic flow provided
by the component thickness and in part
to the coarse grain structure. When hot
thick plate passes through the hot
working rolls, the interior region receives
less hot working and grain refinement
than the near surface layers. As a result
the center of the thick plate has a
structure more nearly similar to that of
the cast ingot from which the plate was
wrought.
Effects of Pressure Vessel Wall
Thickness and Temperature
There is not exact thickness or
accompanying pressure above which
brittle fracture will occur and below
which ductile fracture will occur in
hydrostatic testing of a vessel. However,
caution should be exercised when the wall
thickness is above 40 mm (1.5 in.) and
when the pressure is above 10 MPa
(1.5 × 103 lbf·in.–2). When testing vessels
that fall into this category, it is good, safe
practice to ensure that water used for
hydrostatic testing be at a temperature of
at least 38 °C (100 °F). In addition, no
pressure should be exerted on the vessel
until the wall temperature both inside
and outside is about the same as that of
the pressurizing liquid, usually water. This
precaution has a twofold effect: (1) there
is less chance for the metal to fail in a
brittle manner when the temperature of
the wall of the vessel is close to the
temperature of the contained liquid and
(2) there is less air entrained in the water
at a temperature of at least 38 °C (100 °F).
Minimum Temperature Limit for Leak
Tests of Thick Walled Steel Vessels
When testing vessels with wall thicknesses
above 40 mm (1.5 in.) and pressures less
than 10 MPa (1.5 × 103 lbf·in.–2) and
where vessels are constructed of steels
whose resistance to brittle fracture at low
temperature has not been enhanced, test
temperatures above 18 °C (65 °F) should
be used to minimize the risk of brittle
fracture during the test. Again, the test
pressure should not be applied until the
vessel structure (inside and out) and its
contents are at about the same
temperature.
Procedure for Heading Up
Vessels for Pressure Tests
Before application of pressure within
vessels or systems to be subjected to
pressure tests, it is essential to close and
seal all openings in the vessel pressure
boundary so that pressurizing fluids do
not escape or leak. The operation of
assembly (also known as heading up) of a
vessel for pressure testing must be done
with adequate care to ensure safety. There
are small details, many of which seem
insignificant but could be potential
hazards, that must be given careful
attention. The small details are items that
usually cause most of the problems
because they are the most easily
overlooked. A checklist type of test
procedure is recommended to ensure that
details are not overlooked and that safe
practices are followed.
Safety Aspects of Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
141
Precautions during
Installation of Blind
Flanges and Covers
Precautions in Tightening
Studs or Bolts on Gasketed
Flanges or Covers
The first item to consider when planning
the installation of blind flanges and
covers for opening and open connections
in pressure vessels before pressure testing
or pressure leakage testing is the material
to be used for these closures when
heading up the vessel. Blind flanges and
covers, as well as the bolts or studs with
which they will be attached, must be of
the proper material, thickness and size. If
the cover is burned or sawed from plate,
the outside diameter should have no
notched areas that could serve as points
of stress concentration or nuclei for crack
propagation.
It is important that a flange be tightened
evenly so that equal pressure is applied to
the gasket. Having the flange cocked to
one side by tightening one side more than
another will almost always result in a
gasket blowout.
When bolting up, it is necessary that
every thread on the nut be engaged by a
thread on the bolt or stud or, in
machinist’s terms, one must have a full
nut. Be cautious in situations where the
opening consists of a pad on the vessel
with bottom tapped holes for studs.
Assume that a pad is headed up for testing
and that the studs are of different lengths.
Or perhaps the studs are the same length
and some of them are not threaded far
enough into the tapped holes. A rule of
thumb is that a stud must be threaded in
to a depth equal to or greater than its
diameter. There is only one way to be sure
the threaded attachment is safe. Remove
the bolts or studs from the threaded hole
to see how many threads were engaged.
Many times the person doing the bolting
will not be the one standing beside the
tank watching the gage pressure increase
during the test. This is unfortunate, for
the installer might be much more careful
if he or she expected to be present for the
test. Under no circumstances should a
stud not driven to the proper engagement
depth in blind hole be cut off at the nut
end. Each stud or bolt should be turned
into the threaded hole to provide a
uniform length of threaded connection to
the proper depth. Failure to ensure
adequate depth of engaged threading
results in an unacceptable and highly
dangerous condition that could result in
catastrophic failure when test pressures
are applied.
Precautions in Selection
and Installation of Flange
Bolts for Pressure Tests
Bolting or studding is a vital area for
careful safety conditions. Carbon steel
bolts, studs and nuts are generally
recommended for attaching flange covers
to pressure vessels for working pressures
below 1.7 MPa (250 lbf·in.–2) for tests at
temperatures below 200 °C (400 °F). For
temperatures exceeding 200 °C (400 °F),
alloy steel bolts, studs and nuts are
recommended regardless of the test
pressure. For pressures exceeding 1.7 MPa
(250 lbf·in.–2), only alloy steel studs or
bolting should be used. When alloy steel
studs, bolts and nuts are necessary, their
thread pitch should be not less than
3 mm (eight threads per inch). Loading to
be applied to studs or bolts is
recommended by suppliers of the various
types and sizes of flanges and gaskets. Bolt
or stud loading is generally expressed in
terms of the torque required to tighten a
nut or bolt to give a specific longitudinal
stress (megapascal) in the stud or bolt for
the specific stud or bolt material and cross
section. Also specified by suppliers are the
proper gasket pressures in megapascal or
pound force per square inch. One
important precaution is to determine
whether or not the stud or bolt torque
recommended applies for a lubricated or a
dry bolt or stud. Lubrication of bolt or
stud threads results in a great (and
variable) increase in the actual stud or
bolt stress and in the applied gasket
pressure, for any specified level of stud or
bolt torquing.
142
Leak Testing
Precautions in Selection
and Installation of Gaskets
for Pressure Tests
The choice of gasketing is important and
vital for safety in pressure testing of
vessels and systems with gasketed
attachments, flanges and instrument
connections. In most cases, a soft rubber
gasket may be sufficient for low pressure
testing. However, as the test pressure
increases the gasket strength must be
increased to prevent gasket failure where a
portion of gasket or a small stream of
high pressure liquid may be expelled with
considerable force. For low pressure
testing, a flat elastomer gasket of 60 to
70 durometer will make a safe seal. For
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
higher pressures, one can use the same
type of gasket, reinforced with some type
of fiber. Beyond the range of application
of the elastomeric gaskets, an asbestos or
some other fiber gasket can be used. For
very high pressures, a metal-to-metal seal
such as the ring type joint or other
patented seals must be used.
Proper gasket width and thickness are
important, particularly with high
pressures. With a fixed bolt load, a gasket
that is too wide will result in low gasket
pressure and consequent saturation and
blowout. When a gasket is too narrow, a
high gasket pressure will result and either
the gasket will be crushed to uselessness
or the flange may become grooved or
warped. There are many flange designs on
the market and it is important that
correctly proportioned gaskets of suitable
material and thickness be used to produce
the correct ratio between effective bolt
areas and the gasket contact surface areas.
The suppliers of gasket materials generally
have published data for proper
application of their products. One should
never exceed the recommended operating
conditions for a gasket. A gasket failure
may be a major eye hazard because pieces
of gasket or a jet stream of liquid of gas
under high pressure could cause a serious
eye injury. One should always avoid being
in direct eye line with a gasket while a
vessel is being pressurized. Safety glasses
and face shields should be worn while
inspecting any vessel under pressure. A
good means of protection when
approaching the maximum design
conditions of a gasket would be to wrap
the outside diameter of the flange and
cover with rope and surround the flange
with a secured metal shield of at least
2 mm (0.08 in.) thickness.
Installation and Care of
Sight Glasses for Pressure
Tests
Certain precautions must be observed in
installing and using sight glasses. Sight
glass ends should be cut square and free
of chips, scratches and rough edges. Care
must be taken to protect the glass from
scratches or severe deformation that
might cause failure by explosion. A small
scratch on the surface can greatly weaken
the glass. Deformations caused by objects
bearing against the glass or by improper
tightening of the flange bolts can cause
serious difficulties. It is important that the
temperatures at testing be held constant
or allowed to vary slowly enough to keep
all parts of the sight glass assembly at
approximately the same temperature to
avoid localized stresses in the glass.
Properly designed, applied and installed
safety valves, maintained in good
operating condition, are essential to the
safety of personnel and the protection of
equipment during pressure leak testing
and during abnormal operating
conditions in service. Inspections should
be made of safety valves and overpressure
relief devices to make sure that their
performance meets the requirements of a
given test operation and those for a given
installation in operating equipment.
Functions and Types of
Safety Valves and Pressure
Relieving Devices
Pressure relieving safety devices can be
divided into five basic classifications:
(1) spring loaded devices, (2) weight
loaded devices, (3) pressure loaded
devices, (4) pilot operated devices and
(5) rupture disks. All of these types of
devices are designed to function
automatically at a predetermined set
pressure to prevent excessive
overpressures in the equipment on which
they are installed. The term safety valve is
often used loosely to indicate any or all of
these types of pressure relieving devices.
Normally, safety valves and their
discharge systems are used for pressure
vessels and equipment designed for a
maximum allowable working pressure in
excess of 100 kPa (1 atm or 14.7 lbf·in.–2
absolute). Table 9 lists specifications and
codes applicable to pressure relief devices.
The following terminology and
definitions identify the devices in each of
the preceding categories.
1. Safety valves are automatic spring
loaded pressure relieving devices
actuated by the static pressure
upstream of the valve and are
characterized by rapid full opening or
pop action. Safety valves are used on
steam boilers, drums and super
heaters. They may also be used for
general air, steam and pressurized
gases during service or in leak testing.
2. Relief valves are automatic spring
loaded pressure relieving devices
actuated by the static pressure
upstream of a valve that lifts in
proportion to the increase in pressure
over the operating pressure. Relief
valves are used primarily in systems
filled with liquids.
3. Safety relief valves are automatic
spring loaded pressure relieving
devices actuated by the static pressure
upstream of the valve. They are
characterized by rapid full opening or
pop action on gas or vapors and are
suitable for use either as a safety valve
or as a relief valve, depending on the
application. There are two types of
Safety Aspects of Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
143
safety relief valves. Conventional
safety relief valves are constructed in
such a manner that the back pressure
on the downstream side of the valve
affects the action of the valve.
Balanced safety relief valves have been
balanced by the addition of a pressure
balancing mechanism (bellows, piston
or both) to decrease the valve’s
sensitivity to change in back pressure.
4. Pilot operated safety relief valves are
pressure relief valves in which the
major relieving device is combined
with and is controlled by a
self-actuated pilot relief valve. Pilot
operated safety release valves consist
of two basic units: a pilot or control
unit and the main valve. These two
basic units are mounted either on the
same or on separate connections,
depending on their design. The pilot is
a spring loaded valve that senses the
pressure differential and causes the
main valve to open and close.
5. Pressure and vacuum vents are
automatic pressure or vacuum
relieving devices actuated by the
pressure or vacuum in the protected
vessel or tank. These pressure vacuums
vents fall into two main categories:
weight loaded pallet vents and pilot
operated vents.
6. Rupture disks are thin diaphragms
usually held between special flanges.
They are designed to rupture at a
predetermined pressure so as to relieve
pressure from a vessel or system being
protected.
Terms Related to
Applications of Pressure
Relief Devices
The following terms are related to the
design and application of safety valves
and the pressure systems on which such
valves may be applied.
Maximum allowable working pressure is
defined in the construction codes for
pressure vessels. The maximum allowable
working pressure depends on the type of
material, its thickness and the service
conditions set as the basis of design. The
vessel may not be operated above this
pressure or its equivalent at any metal
temperature other than that used in
specifying its design. Consequently, for
that metal temperature, it is the highest
pressure at which the primary safety valve
is set to open.
The operating pressure of a vessel is the
gage pressure to which the vessel is
usually subjected in service. A processing
vessel is usually designed for a maximum
allowable working pressure that will
provide a suitable margin above the
TABLE 9. Typical specifications and standards for pressure relief devices, including those applicable in petroleum
refineries.
Issuer
Specification or Standard
API
Bulletin 2521, Use of Pressure-Vacuum Vent Valves for Atmospheric Pressure Tanks to Reduce Evaporation Loss
Guide for Inspection of Refinery Equipment: Chapter 5, Preparation of Equipment for Safe Entry
RP 520, Recommended Practice for the Sizing, Selection and Installation of Pressure-Relieving Systems in Refineries
RP 521, Guide for Pressure-Relieving and Depressuring Systems
RP 576, Inspection of Pressure-Relieving Devices
Standard 526, Flanged Steel Pressure Relief Valves
Standard 527, Seat Tightness of Pressure Relief Valves
Standard 620, Design and Construction of Large, Welded, Low-Pressure Storage Tanks
Standard 2000, Venting Atmospheric and Low-Pressure Storage Tanks Nonrefrigerated and Refrigerated
ASME
ASME Boiler and Pressure Vessel Code:
Section I, Power Boilers
Section IV, Heating Boilers
Section VI, Recommended Rules for Care and Operation of Heating Boilers
Section VII, Recommended Rules for Care of Power Boilers
Section VIII, Pressure Vessels
ASTM
A 216, Standard Specification for Steel Castings, Carbon, Suitable for Fusion Welding, for High-Temperature Service
A 217, Standard Specification for Steel Castings, Martensitic Stainless and Alloy, for Pressure-Containing Parts, Suitable for
High-Temperature Service
A 351, Standard Specification for Castings, Austenitic, Austenitic-Ferritic (Duplex), for Pressure-Containing Parts
F 1508, Standard Specification for Angle Style, Pressure Relief Valves for Steam, Gas, and Liquid Services
NBBPVI NB 23, National Board Inspection Code
NB 27, National Board Rules and Recommendations for the Design and Construction of Boiler Blowoff Equipment
144
Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
operating pressure to prevent an
undesirable operation of the safety valve.
Set pressure is the inlet gage pressure at
which the safety valve is adjusted to open
under service conditions. In a liquid
service, the set pressure is the inlet gage
pressure at which the valve starts to
discharge under the service conditions. In
a gas or vapor service, the set pressure is
the inlet gage pressure at which the valve
pops under service conditions.
Cold differential test pressure is the gage
pressure at which the valve is adjusted to
open on the valve or leak test stands. This
cold differential pressure includes the
corrections for service conditions of back
pressure, temperature or both.
Accumulation is the pressure increase
over the maximum allowable working
pressure of the vessel during discharge
through the safety valve. It is expressed in
kPa or lbf·in.–2, or as a percentage of the
maximum allowable working pressure
(MAWP). Maximum allowable
accumulations are established by the
applicable ASME Codes for operating and
fire contingencies.
Overpressure is the pressure increase
over the set pressure of the safety valve. It
is the same as the accumulation when the
safety valve is set at the maximum
allowable working pressure on the vessel.
The overpressure may be greater than the
allowable accumulation if the valve is set
lower than the vessel maximum allowable
working pressure. Likewise, if multiple
safety valves are installed, some with
staggered set pressures above the
maximum allowable working pressure, the
overpressure for the staggered valves will
be less than the allowable accumulation.
Blowdown is the difference between the
set pressure and the reseating pressure of
the safety valve, expressed in kilopascal or
as a percentage of the set pressure.
Lift is the rise of the disk in a safety
valve.
Back pressure is the pressure on the
discharge side of a safety valve.
Superimposed back pressure is the
pressure in the discharge header before
the safety valve opens.
Built up back pressure is the pressure in
the discharge header that develops as a
result of flow after the safety valve opens.
Causes of Improper
Performance of Safety
Valves
Corrosion is one of the basic causes of
difficulties observed in operation of safety
valves. Corrosion may be apparent in
pitting of valve parts, in breaking of
various parts of a valve, in deposits of
corrosive residues that interfere with
operation of moving parts and in general
deterioration of the material in a safety
valve. Leaking valves can allow circulation
of corrosive fluids into the upper parts of
a valve so as to contribute to corrosion of
the movable parts of the valve.
Damaged seating surfaces on safety
valves can contribute to improper safety
valve action during service. API Standard
527-78, Commercial Seat Tightness of Safety
Relief Valves with Metal-to-Metal Seats, gives
acceptable leakage rates.16 Seating surfaces
on safety valve must be maintained to
optical precision. Any imperfection of
these seating surfaces will contribute to
improper valve action in service, as during
leak testing. Foreign particles such as mill
scale, welding spatter, coke or dirt that get
into the valve inlet and pass through the
valve when it opens may destroy the
precision seat contact required for leak
tightness in most safety valves.
Valve chatter causes hammering that
sometimes damages safety valve seating
surfaces severely. Careful handling of the
valve during all phases of maintenance,
installation and disassembly is important.
Bumping or dropping the valve during
installation should be carefully avoided.
all valve parts, particularly guiding
surfaces, should be checked thoroughly
for any type of fouling. Lubrication of all
sliding surfaces with molybdenum
disulfide compounds or graphite and
grease is recommended for safety valve
used in refinery service where valves and
piping can sometimes become plugged by
process solids such as coke and solidified
products.
Causes of Leakage in
Safety Valves
Leakage past the seating surfaces of a
valve after it has been leak tested,
installed and placed in service may be
caused by inadequate maintenance or
installation procedures such as
misalignment of the parts. Leakage could
also result by piping strains resulting from
improper support or by complete lack of
support of discharge piping. This leakage
contributes to seat damage because it
causes erosion or corrosion of the seating
surface and thus progressively aggravates
the leakage problem. Valves subject to
vibration, pulsating loads, low differential
between set and operating pressures and
other circumstances leading to valve
leakage should be inspected and tested
more frequently than valves not operating
under such conditions.
Safety Aspects of Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
145
Testing for Leakage in
Safety Valves
A properly designed test block is
important to facilitate setting and
adjustment of each safety valve during its
inspection and repair. Valve settings are
generally set in the maintenance shop by
using water, air or an inert gas such as
bottled nitrogen as the leak testing
medium. Care should be taken and some
overpressure should be applied to the
valve to be certain that the valve is
opening at the proper set pressure. An
audible leak can otherwise be
misinterpreted as the set pressure of the
valve. In most types of safety valves, a
distinct pop occurs at the set pressure,
making misinterpretation impossible.
Incorrect calibration or lack of calibration
of pressure gages is another frequent cause
of improper valve setting. The pressure
range of the gage used to set valves
should be chosen so that the required set
pressure of the safety valve falls within
the middle third of the range of the
pressure gage.
Safety Valve Inspection Standards
Because of the difficulty in obtaining
absolute leakage tightness in most safety
valves, valve manufacturers use a
commercial leak tightness standard
according to which they manufacture
valves. Subsequent rough handling of the
valve can destroy the commercial
tightness and produce excessive leakage in
the valve after it is placed in service.
Rough handling can occur during
shipment, maintenance or installation of
the valve. Occasionally, safety valve
manufacturers are in a position to assist
the user in establishing inspection and
test intervals for safety valve. Each
manufacturer is familiar with the nature
of the loading, the stress levels and the
operating limitations of their particular
designs, thus enabling them to suggest
inspection intervals appropriate for their
valve equipment. In some instances, the
frequency of inspecting and testing safety
valves used in service is established by
regulatory bodies. This should be
investigated for each locality to avoid any
possible conflict between such regulations
and the frequencies of valve inspection
considered to be satisfactory on some
other basis.
Advantages of Testing Safety Valves
with Air or Nitrogen
Air or inert gas is generally used to test
safety valves, relief valves and safety relief
valves for both set pressure and for
leakage tightness. In general, some means
146
Leak Testing
is required to blind the valve discharge.
Leakage may be detected qualitatively by
placing a thin membrane (such as a wet
paper towel) over the outlet and noting
any bulging of the membrane. A
quantitative measurement can be made by
trapping the leakage and conducting it
through a tube submerged in water, so
that bubble emissions can be observed.
Leaking valves can also often be detected
with ultrasonic leak detectors.
Limitations of Testing Safety Valves
with Water
Testing of safety valves with water is
usually limited to measuring the set
pressure because very small leaks cannot
be readily detected when using water as
the test medium. Water tends to clog
small leaks and prevent detection of
leakage. For this reason, leakage rate and
leak tightness tests of relief valves are
usually made with air as the pressurizing
medium.
Inspection of Safety Valves
on Steam Boilers
Inspection of safety valves on steam
boilers should be carried out in
accordance with local regulatory
requirements as well as in conformity
with manufacturer’s recommendations
and operating company practice. Because
Section I of the ASME Code14 does not
permit block valves between boilers and
boiler safety valves, testing on the
equipment must be done periodically by
raising the steam pressure to pop the
valves while the boiler is in operation.
Precision calibrated pressure gages should
be used during the test procedure. The
accumulation and blowdown should also
be noted. The ASME Code also requires
that the boiler safety valves have a
substantial lifting device by which the
valve disk may be lifted from its seat
when there is at least 75 percent of full
working pressure on the boiler. This
permits checking to be sure that the
moving parts are free to operate.
Frequency of Inspection of
Safety Valves
The inspection of safety valves provides
data that can be evaluated to determine a
safe and economical frequency for
scheduled inspections. This frequency can
be expected to vary greatly because of the
different operating conditions and
environments to which safety valves are
frequently subjected. Usually the intervals
between inspections are increased as a
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
result of satisfactory operating experiences
and are decreased where corrosion,
fouling and leakage problems exist.
Historical records reflecting periodic test
results and service experiences for each
safety valve are valuable for establishing
safe and economical inspection
frequencies.
A definite time interval between
inspections should be established for
every safety valve on operating equipment
to ensure proper performance. The time
interval should be sufficiently firm to
ensure that the inspection is
accomplished but it should be sufficiently
flexible to permit revision and temporary
waiving where justified by circumstances.
The interval between inspections is
normally determined by operating
experience. Obviously, the interval
between inspections of a valve in
corrosive and fouling service conditions
would be shorter than for the same valve
in a clean and nonfouling service.
Where corrosion, fouling and other
service conditions are not known and
cannot be predicted with any degree of
accuracy (as in a new type of process or in
occasional use during leak testing), the
initial inspection should be accomplished
as soon as practical after operations begin
to establish a safe and suitable testing
interval. Safety valves in service should
carry an identifying tag or plate. This
identification is needed to minimize
errors in testing and handling of safety
valves. Identification of safety valves is
essential in keeping accurate historical
records on each valve.
determine the pop pressure of the valve
when removed from service. If the valve
opens at the set pressure, it need not be
tested further to determine the as-received
relieving pressure. If the initial pop is
higher than the set pressure, it is
advisable to make a second test for pop
pressure. If the valve then pops at about
the set pressure, this indicates that the
valve was probably stuck because of
deposits. If the valve does not pop near
the set pressure, this indicates that the
valve setting was higher in error originally
or that it may have been changed during
operation. The as-is test pressure should
be recorded for review and facilitation of
any necessary corrective action.
Routine Checking of Safety Valve Set
Pressure and Leak tightness
The valve parts that most often require
cleaning are the nozzle, springs and seats.
Deposits that are difficult to remove
should be cleaned off with solvents or
wire brushing or should be carefully
scraped. The dismantled parts should be
checked carefully at this time for wear
and corrosion. Checking of valve
components is important. It should be
done carefully with the proper equipment
calibrated for measuring valve dimensions
and with frequent reference to the proper
valve drawings and literature. Parts that
are worn or damaged should be replaced
or reconditioned. Parts such as damaged
springs or bellows should be replaced
without attempting repairs. The valve
body and bonnet may be reconditioned
by means considered suitable for repairs
to other pressure containing parts of
similar materials. After the valve has been
inspected and reconditioned, it should be
assembled in accordance with the
manufacturer’s instructions as to the order
of assembly and the procedure for
adjustment of the various parts.
An important phase of the safety valve
maintenance routine is to determine set
pressure and leak tightness of the valve
both in the as-received condition and
after overhauling. A visual inspection of
safety valves should be made as the valves
are removed from the system or from a
leak testing setup. Many types of deposits
or corrosion products may be loose and
drop out of the safety valve while it is
being transported to the shop for
inspection and repair, if needed. Any
obstruction in the valve should be noted
and corrected. Inspection of the piping or
flange connections at the location of the
safety valve should be done to detect
evidence of corrosion, indications of
thinning and deposits that may interfere
with valve operation.
Determining Safety Valve Pop
Pressure before Dismantling
Before the safety valve is dismantled, it is
generally considered important to
Maintenance Procedures
for Safety Relief Valves
When safety relief valves are to be given
maintenance servicing, each valve should
be carefully dismantled in accordance
with its manufacturer’s instruction
manuals and recommendations. Proper
facilities should be available for
segregating valve parts as the valve is
dismantled. At each stage in the
dismantling process, the valve, stem,
guide, disk, nozzle and other parts require
visual testing. The bellows in balanced
type valves should be checked for cracks
or other failures that might permit leakage
or affect valve performance.
Cleaning, Repair and Replacement of
Safety Valve components
Safety Aspects of Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
147
Setting Repaired Relief
Valves to Required Pop
Pressure
After a used relief valve has been
reconditioned and reassembled, it is ready
for the final spring adjustment to the
required set pressure. The manufacturer’s
recommendations should be used as a
guide in adjusting the spring to the
correct setting. If a new pressure setting is
required, the manufacturer’s limits for
adjustment of the spring must not be
exceeded and applicable ASME Code
requirements must be observed. It may be
necessary to provide a different spring.
After the final adjustment is made, the
valve should be popped at least once to
prove the accuracy of the setting. The
final pop should be within the
manufacturer’s listed accuracy for the cold
set pressure before the valve is approved
for service. Allowance for hot setting
should be made in accordance with the
manufacturer’s data.
Checking Reassembled
Safety Valve for Leak
Tightness
After a reconditioned relief or safety valve
has been satisfactorily checked for
conformance to the set pressure, it is then
desirable to check the valve for leakage.
Excessive leakage could lead to fouled or
inoperable valves, hazard to personnel
and equipment and possible loss of leak
testing fluid or product from processing
systems (see discussion of safety valve
leakage). All necessary records for
inspection, repair, assembly and resetting
should be completed before the valve goes
back into service. These records are
important for effective future use of the
valve. They will provide guidelines for
replacement of valves and components as
well as providing the historical record of
the conditions and services under which
the valve operated.
Need for Keeping
Permanent Records for
Safety Valves
A complete permanent record file should
be kept for each safety valve in service.
The record should provide specification
data for the valve and a history of
inspection and test results. The
specification record is needed to provide
basic information needed to evaluate the
adequacy of the valve for a given leak
testing operation or permanent
148
Leak Testing
installation. It also provides correct
dimensional and material information to
minimize shop errors and expedite
repairs. Historical records showing dates
and results of inspections on safety valves
are necessary for a followup on the
control phase of the program. One of the
foremost reasons for keeping service
records is that they provide a practical
and realistic basis for maintaining safe
and economical inspection intervals that
provide safety to all operators using the
valves.
Precautions with Venting
Devices on Atmospheric
Storage Tanks
Atmospheric storage tanks are widely used
in petrochemical industries. Venting
devices are usually mounted on top of
these tanks to protect the tank from
damage due to excessive internal pressure
or from excessive vacuum. Venting
devices are all to often taken for granted
and forgotten once they have been
installed. They must be considered when
leak testing atmospheric tanks by pressure
change or flow measurements. These
relatively simple venting devices will
normally work properly for long periods
with little attention, but if one fails, it can
result in catastrophic failure of the tank
and loss of its product.
The two main types of venting devices
are breather vents and conservation vents.
Breather vents or open vents usually take
the form of an open pipe of
predetermined size. They permit the
equalization of pressure inside a tank with
the varying external atmospheric pressure.
In general, breather vents are used when
the product stored has a flash point about
40 °C (100 °F) and evaporation losses are
not a concern. The vent should be
equipped with a return bend or weather
head to exclude rainfall, both being
equipped with screens to prohibit any
entry of animals or any other foreign
matter. Vents should be designed so that
any condensate will drain back into the
tank without creating a trap or pocket.
The vent should be located so there is the
least chance of encountering an ignition
source when flammable materials are
stored within the tank. Additionally, the
vent should in no case be smaller than
the discharge or withdrawal connection.
It is bad practice to manifold vents. Each
tank should have its own vent.
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
Conservation Vents or
Breather Valves
Conservation vents or breather valves
isolate a tank until specific pressure or
vacuum levels are reached (relative to
atmospheric pressure). The standard
conservation vent usually relieves at
pressure or vacuum gage levels of 215 Pa
gage (0.5 oz·in.–2 or 0.865 in. of water).
The pressure setting is determined by the
weighting of pallets. The heavier these
weights are, the greater the pressure
difference must be to open them.
Conservation vents can reduce
evaporation losses by 50 percent over
breather vents and each additional
increase of 400 Pa (1 oz·in.–1) in the valve
setting will further reduce the breathing
losses by about 7 percent. These
conservation vents are used where
evaporation loss is a concern and/or when
the product being stored has a flash point
equal to or less than 40 °C (100 °F).
Functions of Vents
With modern welded metal tanks and
roofs, storage tanks have become airtight
vessels. Because of this, it is important to
ensure that the tank has some means of
equalizing the external and internal
pressure. Normal venting devices do not
eliminate evaporation losses but they do
reduce these losses. The majority of
evaporation losses are due to either the
normal tank breathing or to the filling of
the tank. The breathing of the tank refers
to the action caused by increasing
atmospheric temperature or decreasing
atmospheric pressure. The increasing
temperature causes the vapor pressure of
the tank to increase until it is greater than
the atmospheric pressure and the vapors
of the tank are driven out until the
pressure is equalized. The normal
breathing cycle involves exhaling during
the late morning and early afternoon and
inhaling during the evening when the
temperature decreases. Likewise, as
atmospheric pressure decreases, the vapor
pressure inside the tank becomes greater
than the surrounding atmosphere and the
vapors of the tank are driven out until the
pressure is equalized. When the tank is
being filled, the liquid coming in acts to
displace the vapors in the tank, causing
these vapors to be driven out. Both
actions would cause a differential pressure
far in excess of the normal design pressure
of atmospheric tanks if the movement of
vapors were prohibited and the tank acted
as a closed system.
The vent reduces the evaporation losses
by adding another resistance to the
normal vapor movement; it does not
prohibit the movement. The effect is that
a pressure slightly below the design
pressure is maintained on the tank. It
makes it harder for the vapors to escape.
The resistance is caused by an orifice
effect in breather vents and by the
pressure setting (the pressure required to
open the pallets) on conservation vents.
Beside reducing evaporation losses, the
vent is also a safety device. The safety
aspect has priority over loss reduction.
Safety is the first concern when selecting
the proper vent. Other considerations
necessary when determining the proper
vent are filling and emptying rates for the
tank, the size of the tank, the product
being stored, the strength of the tank and
the normal daily ambient temperature
change rates.
Effects of Flame Arrestors
in Vents
Flame arrestors consist of a group of
tightly spaced metal plates placed at the
entrance to a vent. They are intended to
prevent a flashback of flame through a
vent, which could cause an explosion of
flammable products in a tank. For a
flashback to occur, an ignition source
must be present and the tank must be
expelling flammable vapors. The theory of
the flame arrestors is that they should
dissipate enough of the heat energy to
prevent a flame front from passing
through them. However, many users now
believe that a conservation vent will
prohibit flashback just as well as a flame
arrestor without the maintenance
problems caused by a flame arrestor.
Therefore, a tight steel roof and a
conservation vent may provide all the
protection that is required. The negligible
additional protection offered by a flame
arrestor may not warrant assuming the
maintenance problems and risk of tank
damage as a result of a flame arrestor
clogging up or prohibiting flow. This topic
is discussed in Petroleum Safety Data
Publication PSD 2210, Flame Arrestors for
Vents of Tanks Storing Petroleum Products,17
compiled by the Committee on Safety and
Fire Protection of the American Petroleum
Institute.
Safety Aspects of Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
149
PART 9. Exposure to Toxic Substances
The following discussion pertains to
measurement and reporting of
recommended limits of exposure to toxic
substances by United States government
agencies.
Threshold Limit Value and
Time Weighted Average
The threshold limit value (TLV) is a
recommended upper limit (ceiling) or
time weighted average (TWA)
concentration of a substance to which
most workers can be exposed without
adverse effect. This concentration may be
designated as a ceiling (C1) or time
weighted average (TWA) concentration.
The notation (SKIN) indicates that even
though the air concentration may be
below the limit value, significant
additional exposure to the skin may be
dangerous. Threshold limit values are
quantified in TLVs: Threshold Limit Values
for Chemical Substances and Physical Agents
in the Work Environment, (third edition,
1971), its supplement or from
documentation in the annual reports of
the America Conference of Governmental
Industrial Hygienists (ACGIH).18
NIOSH Water Quality
Toxicity Ratings
The National Institute for Occupational
Safety and Health Aquatic Toxicity ratings
are published in Water Quality
Characteristics of Hazardous Materials.19
The format for this line is AQUATIC
TOXICITY RATING: Tlm96 µL·L–1 where
TLm96 is defined as the 96 h static or
continuous flow standard protocol.
Because of the lack of standardization and
the wide variety of species investigated,
ratings are used to give an indication of
the toxicity of substances to aquatic life.
material is properly classed, described,
packaged, marked, labeled, and in the
condition for shipment as specified by
49 CFR, Parts 100 to 189. For
transportation purposes, a hazardous
material means a substance or material
which has been determined by the
Secretary of Transportation to be capable
of posing an unreasonable risk to health,
safety, and property when transported in
commerce and which has been so
designated.
Basic hazard classes include compressed
gases, flammables, oxidizers, corrosives,
explosives, radioactive materials, and
poisons. Although a material may be
designated by only one hazard class,
additional hazards may be indicated by
adding labels or by other means.
It is essential, therefore, that all
required labels(s) as well as the hazard
class be known. Generally, poison must
always be labeled as a poison regardless of
the other labeling requirements in order
that adherence to the prohibition against
shipping poisons with foodstuffs can be
assured.
Specific shipping names are designated
for hazardous materials in regulations
because of the presence of many
nontechnical names or the use of archaic
names for some materials.
Determination of the correct
classification for transportation of
materials is the responsibility of the
shipper.
National Institute for Occupational
Safety and Health criteria documents
recommending environmental
(occupational) exposures are currently
available for various toxic substances
encountered in leak testing. The reference
citation (NTIS) is the National Technical
Information Service, United States
Department of Commerce, from which
these publications are available.
Occupational Diseases
Hazardous Substances
Except as provided for certain export and
import shipments, no person may offer or
accept a hazardous material, as defined by
the Code of Federal Regulations [CFR],1
Title 49, for transportation in commerce
within the United States unless that
150
Leak Testing
The National Institute for Occupational
Safety and Health publication
Occupational Diseases — A Guide to Their
Recognition20 (revised periodically)
describes both biological hazards and
chemical hazards and the harmful health
effects of many substances used in
industry. Most of the known occupational
disease producing chemicals are listed by
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
chemical groups, e.g., aliphatic
hydrocarbons, alcohols, glycols. Listed
also are occupations in which workers are
potentially exposed to each toxic agent.
Whether the exposure to the toxic agent
constitutes a hazard depends on such
factors as the concentration of the agent,
how the agent is handled and used,
duration of exposure, susceptibility of the
worker to the agent and health protection
practices adopted by management. Thus,
all hazardous situations imply an
exposure but not all exposures are
hazardous.
Topics covered for each substance or
group of toxic chemicals include the
following: (1) description and chemical
formula, (2) synonyms and common
names for material, (3) potential
mechanisms of occupational exposures,
industries in which exposures can occur
and worker occupations which may lead
to exposures, (4) permissible exposure
limits (if established), (5) routes of entry
of toxic chemical into human body,
(6) harmful effects of toxic substance,
(7) symptoms and systemic effects of
exposure, (8) medical surveillance
recommendations, (9) special tests used or
recommended to detect worker ingestion
or response to toxic chemicals,
(10) personal protective methods and
(11) bibliography of pertinent references.
General warnings are given in other
sections of this book, where experience
indicates that possible hazards may exist.
However, this volume is devoted to leak
testing; its users are referred to qualified
authorities on industrial safety, toxic
substances, exposure limits, biological
effects, and legal requirements and
responsibilities. For advice, the user
should refer specifically to plant safety
rules and procedures; local, municipal,
county, state and national laws and
regulations; and qualified safety and
health organizations and agencies.
Safety Aspects of Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
151
References
1. Code of Federal Regulations.
Washington, DC: United States
Government Printing Office.
2. Hemeon, W.E. Plant and Process
Ventilation. New York, NY: Industrial
Press (1963).
3. Roehrs, R.J. and D.E. Center. “The
Safety Aspects of Leak Testing.” ASNT
Fall Conference [Detroit, MI, October
1968]. Abstract in Materials Evaluation,
Vol. 26, No. 9. Columbus, OH:
American Society for Nondestructive
Testing (September 1968): p 34A.
4. Nondestructive Testing Handbook,
second edition: Vol. 1, Leak Testing.
Columbus, OH: American Society for
Nondestructive Testing (1982).
5. Hine, C.H. and N.W. Jacobson. “Safe
Handling Procedures for Compounds
Developed by the Petro-Chemical
Industry.” AIHA Journal. Vol. 15.
Fairfax, VA: American Industrial
Hygiene Association (June 1954):
p 141-144.
6. NIOSH Registry of Toxic Effects of
Chemical Substances. HEW Publication
NIOSH 78-104A. Washington, DC:
United States Department of Health,
Education and Welfare (1978).
7. NFPA 77, Recommended Practice on
Static Electricity. Quincy, MA: National
Fire Protection Association (1993).
8. ASTM D 396, Specification for Fuel Oils.
West Conshohocken, PA: American
Society for Testing and Materials
(1980).
9. ASTM D 323, Test Method for Vapor
Pressure of Petroleum Products (Reid
Method). West Conshohocken, PA:
American Society for Testing and
Materials (1982).
10. National Electrical Code. Quincy, MA:
National Fire Protection Association
(1996).
11. Holler, L. R. Ultraviolet Radiation. New
York, NY: John Wiley & Sons (1952).
12. Criteria for a Recommended Standard for
Occupational Exposure to Ultraviolet
Radiation. USGPO No. 1733-000-12.
Washington, DC: United States
Government Printing Office.
13. NFPA 51, Standard for the Design and
Installation of Oxygen-Fuel Gas Systems
for Welding, Cutting, and Allied
Processes. Quincy, MA: National Fire
Protection Association (1997).
152
Leak Testing
14. ASME Boiler and Pressure Vessel Code.
New York, NY: American Society of
Mechanical Engineers.
15. SE 432-95, Standard Recommended
Guide for the Selection of a Leak Testing
Method [ASTM E 432-71 (1984)]. New
York, NY: American Society of
Mechanical Engineers (1995).
16. API Standard 527-78, Commercial Seat
Tightness of Safety Relief Valves with
Metal-to-Metal Seats. Washington, DC:
American Petroleum Institute (1978).
17. American Petroleum Institute,
Committee on Safety and Fire
Protection. Petroleum Safety Data
Publication 2210, Flame Arrestors for
Tank Vents. Washington, DC:
American Petroleum Institute (May
1971).
18. America Conference of Governmental
Industrial Hygienists. TLVs: Threshold
Limit Values for Chemical Substances
and Physical Agents in the Work
Environment with Intended Changes for
1983-84. Cincinnati, OH: American
Conference of Governmental
Industrial Hygienists.
19. Hahn, W. and P. Jensen. Water Quality
Characteristics of Hazardous Materials.
College Station, TX: Texas A&M
University (1974).
20. Key, M.M. Occupational Diseases — A
Guide to Their Recognition. DHEW
publication NIOSH 77-181.
Washington, DC: United States
Department of Health, Education, and
Welfare [DHEW], National Institute for
Occupational Safety and Health
[NIOSH]; Superintendent of
Documents, United States
Government Printing Office (1977).
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
C
5
H A P T E R
Pressure Change and Flow
Rate Techniques for
Determining Leakage Rates
Charles N. Sherlock, Willis, Texas
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
PART 1. Introduction to Pressure
Instrumentation, Measurements and Analysis
Functions of Pressurizing
Gases in Leak Testing
Atmospheric air and nitrogen are often
used as pressurizing fluids in leak testing
and leakage measurements. Their fluid
pressure serves to create pressure
differentials across pressure barriers or
walls. This pressure differential, in turn,
causes the pressurizing gas to flow, by
various mechanisms, through leaks in the
containment walls.
Leaks are the physical holes or
passageways that may exist in wall
materials, welds, mechanical seals or
joints. The fluid that flows through the
leak passageways constitutes leakage. The
rate of leakage in turn is taken as a
measure of the size of the leak.
In general, the higher the differential
pressure, the greater the rate of leakage.
With higher rates of leakage, the
sensitivity of leak detection and leakage
measurement is typically increased.
Closed systems with air or other gas
pressures above atmospheric pressure
(101.325 kPa) respond to leakage by
pressure changes (within closed systems)
or require inflow of gas to maintain
constant pressure conditions. These
pressure changes or rates of fluid flow can
be used to determine (1) the presence of
leaks or (2) the rates of leakage, when
internal volumes, fluid temperatures and
other variables are known or can be
measured accurately. The physical
properties and characteristics of the
pressurizing fluids must be known and
the effects of fluid reactions to various test
conditions must be calculated to make
quantitative measurements of leakage
rates. Pressurizing gases should obey the
ideal gas laws. In some cases, the effects of
water vapor and other gaseous materials
that do not obey the general gas laws
must be determined and their effects
subtracted from the pressure
measurements.
Compressibility of Gaseous
and Liquid Fluids
Gases are frequently regarded as
compressible and liquids as
incompressible. Strictly speaking, all fluids
are compressible to some extent.
Although air is usually treated as a
compressible fluid, there are some cases of
flow in which the pressure and density
changes are so small that the air may be
assumed to be incompressible. Examples
include the flow of air in ventilating
systems and the flow of air around aircraft
at low speeds. Liquids like oil and water
TABLE 1. Typical operating ranges and probable accuracy limits of pressure gaging systems.
Pressure Measuring
Instruments
Deadweight testing machines
with various operating ranges
Mechanical dial pressure gages
Quartz Bourdon tube gages
Metal Bourdon tube gages
Water U-tube manometer
Direct-reading mercury manometer
Digital U-tube mercury manometer
Digital aneroid capsule
Ion mass detector sensor
Ranges of Pressures
________________________________________
SI Unit (kPa)
2 to 350
350 to 3500
3500 to 16 000
16 to 80 000
0 to 700 000
0 to 20 000
7000 to 140 000
0 to 7.5
0 to 350
0 to 285
35 to 3500
50 to 800
English Units
Accuracy Limits
_____________________________________________
SI Units
(0.3 to 50 lb f ·in.–2)
typically about 0.003 percenta
–2
(50 to 500 lb f ·in. )
typically about 0.003 percenta
–2
(500 to 2400 lb f ·in. )
typically about 0.003 percenta
(2400 to 12 000 lb f ·in.–2)
typically about 0.003 percenta
(0 to 100 000 lb f ·in.–2)
±0.066 to ±2 percent of full scale
(0 to 3000 lb f ·in.–2)
±0.01 to ±0.02 percent of full scale
(1 × 103 to 2 × 104 lb f ·in.–2) see manufacturer’s specifications
(0 to 30 in. H2O)
±1 Pa
(0 to 100 in. Hg)
±80 Pa
(0 to 84 in. Hg)
±3 Pa
(5 to 500 lb f ·in.–2)
±0.05 percent of full scale
(7 to 120 lb f ·in.–2 gage)
10 –5 Pa·m3·s –1 leakage rates
English Units
(±0.03 torr)
(±2.5 torr)
(±0.1 torr)
(10–6 std cm3·s–1)
a. Traceable to US National Institute of Standards and Technology.
154
Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
may be considered as incompressible in
many cases; in other cases, the
compressibility of such liquids is
important. For instance, common
experience shows that sound waves travel
through water and other liquids; such
pressure waves depend on the
compressibility or elasticity of the liquid.
Instrument Systems for
Precise Pressure
Measurements during Leak
Tests
Quantitative and reproducible leakage rate
testing by pressure change measurements
depends critically on the control and
measurement of test pressures applied to
systems under test. The most precise
pressure measuring instruments are
deadweight testers. These are used most
commonly only for calibrations of other
pressure measuring instruments. Water or
mercury manometers (U-tubes partially
filled with liquid) are also used for
calibration of other pressure gages and
instruments. Other pressure measuring
instruments include Bourdon gages; rapid
response electrical output signal sensors
used in potentiometric, capacitance,
reluctance and piezoelectric pressure
gages; spiral wound quartz crystal and
wire resistance strain gages; and
specialized electronic gages with digital
output signals of pressure. Table 1 lists
typical pressure gages used in leak testing
of pressurized systems and indicates their
typical pressure range and accuracies.
Deadweight Piston
Calibration Standards for
Pressure Measurements
The deadweight piston gage is a
calibration standard for measuring
pressures. Pressure or force per unit area is
provided by known weights acting on the
known area of the cylinder. Fluid pressure
to be measured is applied against the
bottom of the piston, developing enough
force to lift the weights. Thus, the two
factors of primary importance are the
weights used and the effective area of the
piston-and-cylinder combination.
Figures 1 and 2 show a deadweight
calibration machine.
Three types of deadweight piston gage
are available: (1) simple piston pressure
gage, (2) controlled clearance piston
pressure gage and (3) reentrant piston
pressure gage. The first is simple and most
commonly used. The controlled clearance
FIGURE 1. Schematic of dead weight machine for calibrating force measurement devices.
16 mm (0.63 in.) diameter
hole in stage and lower yoke
From 0 to 450 mm
(0 to 18 in.)
Adjustable loading stage
From 0 to 450 mm
(0 to 18 in.)
Lower yoke
Loading stage
adjustment wheel
Lower pull rod
0.8 m
(31 in.)
200 mm (8 in.) clearance
between yoke tension rods
Yoke assembly (weighed to
0.003 percent accuracy)
Lever to apply
yoke assembly
Yoke assembly
weight rod
0.45, 0.9 and 2.3 kg (1, 2 and 5 lb)
weights applied and removed as
required
Levers to apply weights
All weights smaller than 10 lbf, 2 kgf and 50 N
are applied and removed as required
1.2 m
(46 in.)
Dead weights
Weight supports
Adjustable feet
for leveling
740 mm (29.0 in.)
650 mm (25.5 in.)
Pressure Change and Flow Rate Techniques for Determining Leakage Rates
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
155
gage reduces errors caused by deformation
of the cylinder because of the pressure in
the cylinder. The reentrant gage is a
compromise between the first two types
of gages.
Factors Influencing Piston Gage
Pressure Measurements
Temperature affects the dimensions of
gage piston and cylinder. Gravitational
force depends on the location of the
instrument on the earth’s surface and on
its altitude above sea level. Air is a fluid
and has a buoyant effect on objects
(weights) submerged in it. Compressibility
affects fluid density, which can affect
calibrations if the pressure is measured at
a level different from that of the piston
FIGURE 2. Dead weight machine for
calibrating force measurement devices.
face. All these effects are predictable and
correction factors can be obtained from
the various pressure gage manufacturers,
the National Institute of Standards and
Technology and the local weather bureau.
Measuring Fluid Pressure
with Manometers
The manometer balances hydrostatic
pressures with the weight of a column of
liquid. Thus, the accuracy with which a
pressure can be measured by a manometer
depends on (1) the several factors that
affect the weight of the fluid columns and
(2) the accuracy with which the column
heights can be observed. For the basic U
manometer configuration, if both ends of
the U-tube are open to the atmosphere,
the same pressure acts on each side. Then
the column of liquid on one side of the
U-tube will exactly balance the column of
liquid on the other side. The top surfaces
of the two columns will be at the same
level. However, if one leg of the
manometer is subjected to a pressure
greater than that applied to the other leg,
the heights of the two liquid columns will
differ. The difference in column heights
will be proportional to, and a true
measure of, the differences in pressure
applied to the tops of the liquid columns
in the two legs of the manometer.
The difference in the height of the
liquid in the two legs is exactly the same
whether (1) the diameter of the glass tube
is the same in both legs or (2) the legs
have different diameters, provided that
the diameter of the smaller tube does not
approach capillary diameters where
surface tension effects have an influence
on the height of the liquid columns. The
mercury barometer is an example of a
well type absolute manometer, where
atmospheric pressure operates on the
liquid in the open dish of the well
whereas vacuum pressure acts on the top
of the liquid column in the closed
barometer tube.
Effect of Fluid Density in
Manometers
When a manometer measures a pressure,
the difference in the U-tube liquid
column heights depends not only on the
external pressures applied to the two sides
of the U-tube but also varies with the
density (mass per unit volume) of the
liquid within the U-tube. To illustrate,
suppose that three U-tube manometers
contain oil, water and mercury,
respectively, as their fluids. The difference
in fluid column heights will differ in these
manometers when subjected to the same
differential pressure. The largest difference
in column heights is observed with the
156
Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
low density oil, slightly less with water
and considerably less with the high
density mercury. The differential heights
vary in a ratio of approximately 17 (oil),
14 (water) and 1 (mercury).
Silicon Based Pressure
Sensors
Because of its role in the production of
electronic integrated circuits, enormous
research effort has been committed to
understand, control and commercialize
the electronic performance of silicon as a
semiconductor. A byproduct of research
has been an increased use of silicon as the
sensing member for many types of
electromechanical sensors.
Silicon’s unique mechanical and
electrical properties make it well suited for
sensing various phenomena. Some of
these properties are the following.
1. It has a strength-to-weight ratio five
times greater than stainless steel.
2. It is as hard as quartz.
3. Its thermal conductivity is close to
that of aluminum.
4. It has almost perfect elasticity,
exhibiting no mechanical hysteresis.
5. It is readily machined both
mechanically and chemically to
achieve a required shape or profile.
Furthermore, silicon responds to light,
magnetic fields, stress and temperature
and is impervious to most media. Just as
in its use as a semiconductor, it can be
produced as pure single crystal silicon or
it can be doped with various impurities to
provide specific effects. When a
micromachined silicon chip is used as a
sensor it is a practical matter to include
signal enhancing circuitry directly on the
sensing element just as in integrated
circuits.
temperature compensated to produce a
pressure reading with a high degree of
precision over an extended temperature
range.
A digital pressure transducer (Fig. 3a)
uses a silicon pressure transducer to
provide pressure measurements with an
accuracy of 0.01 percent of full scale over
a temperature range of 15 to 45 °C (59 to
113 °F). The gage is available as absolute,
bidirectional, compound, gage and
vacuum types in full scale ranges as low as
0 to 2.5 kPa (0 to 10 in. H2O) and up to 0
to 41 MPa (0 to 6 × 103 lbf·in.–2). Several
of these digital pressure transducers with
different ranges can be connected in series
for multiple test pressure ranges.
Another pressure instrument (Fig. 3b)
uses a silicon pressure transducer and
handles pressure ranges up to 0 to 70 MPa
(0 to 1 × 104 lbf·in.–2). Standard accuracy
is typically 0.025 percent of full scale,
with temperature compensation of 15 to
45 °C (59 to 113 °F).
A specialized variation of the digital
pressure transducer is the precision
barometer (Fig. 4a). This absolute device
FIGURE 3. Digital pressure gage system components:
(a) digital pressure transducer; (b) console.
(a)
Precision Pressure Measurements
with Silicon Pressure Transducers
Some of the advantages of using a silicon
chip as a strain gage or sensor for
precision pressure measurements are the
following.
1. It can be very small, which reduces
package size.
2. It is highly stable for long term
reliability.
3. It is inherently rugged, so it is
practically immune to the effects of
tilt and vibration.
When a silicon pressure sensor
incorporates temperature sensing the
device can be characterized for pressure
response over a range of temperatures.
Using microprocessors, the unique
pressure/temperature characterization for
an individual silicon sensor can be
(b)
Pressure Change and Flow Rate Techniques for Determining Leakage Rates
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
157
has a fixed pressure range of 75 kPa to
115 kPa (22 to 34 in. Hg) with a
resolution of 0.34 Pa (1 × 10–4 in. Hg).
A practical instrument for testing or
calibrating pressure devices of many
different ranges is the multiple range
pressure standard (Fig. 4b). This
instrument incorporates from four to
seven precision silicon pressure
transducers inside a single chassis with a
common central processing unit and user
interface. Each transducer can be custom
made to operate in any range from 0 to
2.5 kPa (0 to 10 in. H2O) up to 6.9 MPa
(0 to 1 × 103 lbf·in.–2), each with an
accuracy of 0.01 percent of full scale over
the temperature range of 15 to 45 °C (59
to 113 °F). Each transducer is individually
protected from overpressure by relief and
shutoff valves. This system can be
switched between range hold and
autorange. In the range hold mode all tests
are performed using a single range
transducer, whereas in autorange mode the
applied pressure is automatically directed
to the internal transducer that will
provide the highest level of accuracy for
that pressure. In this mode operator or
programmed switching between tests of
different pressure ranges is eliminated.
Precision Regulated Pressure
Output
A pressure calibration system (Fig. 5a)
finds application when precise pressure
output in the range of 0 to 10.35 MPa
(0 to 1.5 × 103 lbf·in.–2) is required. The
pressure calibration system has a
measurement accuracy of up to
0.01 percent of full scale and a 0.002
percent of full scale control stability over
the compensated temperature range of 15
to 45 °C (59 to 113 °F). The pneumatics
module consists of from one to three
internal silicon pressure transducers, the
reed valve regulator, the valves and
plumbing. The system’s macro capability
lets the user program up to 64 different
test routines with up to 256 steps in each
routine.
A high pressure control unit can
extend the range of the pressure
calibration system for precision regulated
pressure up to 40 MPa (6 × 103 lbf·in.–2).
Both units are operated from the front
panel of the pressure calibration system or
FIGURE 5. Pressure measurement instrumentation: (a)
pressure calibration unit; (b) portable pressure standard.
(a)
FIGURE 4. Pressure measurement instrumentation:
(a) barometric pressure gage; (b) multiple range pressure
standard for calibrating pressure transducers.
(a)
(b)
(b)
158
Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
its communication ports. The high
pressure control unit also uses a fully
temperature compensated silicon pressure
transducer with an accuracy of up to
0.01 percent of full scale. Control stability
of the high pressure control unit is better
than 0.01 percent of full scale.
Dual Range Precision Pressure
Measurement in the Field
A field pressure standard (Fig. 5b) is suited
for high accuracy pressure measurement
requiring two different pressure standards
or pressure types. Both pressure ranges use
temperature compensated silicon pressure
transducers, available for pressures from
0 to 2.5 kPa (0 to 10 in. H2O), up to
40 MPa (6 × 103 lbf·in.–2), and up to
0.01 percent of full scale accuracy. Either
pressure range is available with an
absolute, gage, compound, vacuum or
bidirectional pressure transducer.
Digital Pressure Gages
A representative digital pressure gage
(Fig. 6) has as its pressure sensing element
a piezoresistive, strain gage transducer
coupled to solid state circuitry. The
transducer’s integrated strain gage bridge
is diffused on one side of a single-crystal
silicon diaphragm. Application of the
pressure to be measured activates the
silicon diaphragm only slightly. Minimum
movement causes the strain gage bridge
fused to the diaphragm to produce an
electrical signal proportional to the
pressure.
Because there is no mechanical load on
the sensing element, there are no friction
errors. The transducer’s direct current
output is proportional to pressure and is
electronically linearized and compensated
for temperature and line voltage effects. It
is then scaled, stabilized and converted for
high resolution display. The analog
FIGURE 6. Digital pressure gage: (a) photograph; (b) schematic.
(a)
(b)
Analog output
(direct current)
P
Pressure
transducer
Temperature
115/230 V,
50/60 Hz
Ranging
network
Amplifier
Power
supply
Compensation
Voltage
reference
Analog-to-digital
converter
Light emitting diode
display
Pressure Change and Flow Rate Techniques for Determining Leakage Rates
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
159
voltage output can be used for remote
readout or as a process control input.
For pressures up to 1 MPa
(150 lbf·in.–2), differential and gage
measurements are handled by the same
instrument. Gages are available for
pressure ranges between 0 to 0.035 and 0
to 6.2 MPa (0 to 5 and 0 to 900 lbf·in.–2).
FIGURE 8. Absolute pressure dial gage: (a) with 150 mm
(6.0 in.) diameter dial, scale length of 0.75 m (30 in.),
accuracy of 0.1 percent of full scale and full scale ranges
from 7 to 3 400 kPa (1 to 500 lbf·in.–2); (b) typical aneroid
capsule pointer operating mechanism; (c) typical Bourdon
tube pointer operating mechanism.
(a)
Digital U-Tube Mercury
Manometer Pressure
Measurement System
A high precision digital mercury U-tube
manometer system can be used for
measurement and transmission of
pressure readings as binary coded digital
signals to computers, digital display
systems or electronic data processing
equipment. This instrument uses the
principles of ultrasonic pulse reflection for
measurement of transit time of pulses
reflected off the mercury meniscus in each
leg of the manometer. Its sensitivity is
better than 0.3 Pa (2.5 mtorr). Accuracy is
about 3 Pa (25 mtorr) and the direct
reading electronic display is readable to
this accuracy. The pressure ranges of this
instrument extend from 0 to 280 kPa (0 to
2.1 ktorr).
(b)
Pointer
Capsule
stop
Capsule
Precision Calibrated
Absolute Pressure Dial
Gages
A series of precision dial gages are
available for measurement of absolute
Calibration
adjustment
Backlash
eliminator
Pinion
Geared sector
Revolution
indicator
FIGURE 7. Example of two-revolution extended scale precision
dial gage for measuring absolute pressure, calibrated by
methods traceable to National Institute of Standards and
Technology.
Flexure
(c)
Backlash
eliminator
Push rod
Jewel
bearing
Flexures
Reference
Bourdon
Stop
Pointer
Revolution
indicator
Ratio
linkage
Calibration
adjustment
Geared
sector
160
Leak Testing
Pressure
Bourdon
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
pressures with accuracies of 0.066 percent
of full scale readings. Aneroid capsules are
used for the lower pressure range dial
gages; Bourdon tubes are used for the
higher pressure range dial gages. In the
capsule types, pressure is applied to the
case of the gage, which is rated for gage
pressure of 240 kPa (35 lbf·in.–2). The gage
case is also provided with a tempered glass
dial cover and an overpressure blowout
plug on the back of the case. In other
models with a double revolution scale,
accuracy is 0.1 percent of full scale.
Sensitivity is 0.01 percent of full scale and
repeatability is 0.03 percent of full scale.
Gages are aneroid capsule types in
absolute pressure ranges up to 350 kPa
(50 lbf·in.–2). Above 350 kPa, gages
incorporate Bourdon tubes. Bourdon tube
gages have a high strength plastic dial
cover and a blowout plug in the back of
the case. These dial gages are calibrated
with precision mercury manometers or
primary standard pneumatic piston gages,
to provide calibrations traceable to the
National Institute of Standards and
Technology. The gage of Fig. 7 has a dial
diameter of 220 mm (8.7 in.) and a scale
length of 1.15 m (45 in.) and can be
configured for pressure ranges up to
3.5 MPa (500 lbf·in.–2 absolute).
Practical Visual Pressure
Indicators for Leak Testing
in the Shop or Field
Types of pressure gages that provide
visible indications during leak testing
include absolute pressure dial gages
(Fig. 8), aneroid barometers, calibration
instruments (Fig. 9) and ordinary dial
gages indicating pressure relative to
ambient atmospheric pressure (gage
pressure) (Fig. 10), as well as water
manometers, U-tube mercury manometers
and mercury column barometers. The
ordinary calibrated pressure dial gage is
the type used for short duration pressure
hold tests of test channel zones, double
gasket flange interspaces and airlocks. For
short duration pressure hold test,
barometric pressure variations are ignored
FIGURE 9. Electropneumatic calibrator for field applications.
Pressure Change and Flow Rate Techniques for Determining Leakage Rates
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
161
and ordinary pressure gages showing gage
pressures are used. One reason for this
procedure is economic; absolute pressure
gages cost five or more times as much as
ordinary pressure gages. The ordinary dial
pressure gage has typical accuracies in the
range of ±0.25 to ±0.33 percent of full
scale indication when recently calibrated.
A mirror reflector behind the pointer of
quality dial gages permits the observer to
reduce the parallax error in the readings.
When reading a pressure dial gage,
manometer column or quartz manometer,
observers should position their heads so
that their eyes are at the same level as the
indicator on the gage or the top of the
fluid column in the manometer. If the
height of the gage or manometer is other
than normal eye level, the observer
should position the line of sight directly
in front of the gage or manometer. These
pressure readings should not be taken
while viewing at an angle other than
perpendicular to the face of the
instrument. Following these procedures
will help to reduce the variable parallax
error in reading which results when
different observers read test data from the
same instruments. With dial gages
equipped with a mirror reflector, the
reading is taken by aligning the pointer
directly over its own reflection in the
mirror. The same techniques are used
when reading an aneroid barometer
because the barometer is itself an absolute
pressure dial gage.
FIGURE 10. Ordinary dial pressure gage that measures gage
pressure (the difference between actual pressure and
atmospheric pressure), calibrated in inches of water for low
pressure differentials.
Technique for Precision Reading
of Height of Manometer Columns
The reading point for mercury
manometers is the top of the meniscus, as
shown in Fig. 11a. The readings point for
water manometers is the bottom of the
meniscus, as shown in Fig. 11b. When a
manometer is equipped with a mirror
reflector, the reading is taken by aligning
the reading point on the meniscus
directly over the reflection of the reading
point in the mirror.
Techniques for Reading Pressure
Test Instruments Consistently and
Accurately
When reading pressure gages, manometers
or temperature instruments during
absolute pressure leak tests, the operator
should estimate pointer position or meter
indications to at least one half of the
smallest scale division on the instrument.
It is important that the leak testing
operator be consistent in the reading of
instruments. Consistency in reading is as
important as the assurance that the
instrument is properly calibrated. This
consideration results from the
cancellation of calibration errors during
successive instrument readings. For
example, assume that a pressure gage
reads high by 5 kPa at the test pressure. In
this case the initial pressure reading may
be shown as 340 kPa instead of the true
value of 335 kPa. The final reading would
appear as 334 kPa instead of its true value
of 329 kPa. Assuming that the test system
FIGURE 11. Reading points for liquid column
pressure gages, manometers and barometers:
(a) mercury manometer meniscus reading
point; (b) water manometer meniscus
reading point.
(b)
(a)
Reading
point
Reading
point
Mercury
162
Leak Testing
Water
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
remains at uniform temperature, the
pressure loss is found to be 6 kPa
(0.9 lbf·in.–2) in the operator’s gage
readings, as well as in the true pressure
readings.
Variations in Atmospheric Pressure
at Earth’s Surface
All pressure measurements made within
the earth’s atmosphere are influenced by
the fact that the earth’s atmosphere
imposes a pressure on any object in it.
This atmospheric pressure varies not only
with elevation and altitude but also with
time and temperature. Although the
atmospheric pressure at any one location
is not constant, a standard atmosphere is
now specified to be a pressure of
101.325 kPa (14.696 lbf·in.–2 or
760.000 torr). Note that a pressure gage
that indicates 50 kPa gage pressure is
indicating 50 plus 101 or a total of
151 kPa of absolute pressure. Therefore,
an absolute pressure gage would indicate
the absolute pressure of 151 kPa when it
is connected to a source of 50 kPa gage
pressure above atmospheric pressure.
sensing transducer, available in pressure
ranges varying from 10 kPa to as high as
140 MPa (2 × 104 lbf·in.–2). For most
applications, the electronic memory type
of pressure decay leak testing system
permits detection of smaller changes in
pressure and faster testing times can be
obtained. It also eliminates the problems
associated with use of a reference pressure
FIGURE 12. Pressure decay leak tester with pressure sensitivity
of 0.05 percent of full scale, pressure transducers ranging
from vacuum to 140 MPa gage (2 × 104 lbf·in.–2) and full
scale ratings with electronic memory and automatic control
of pressure sensitivity range, delay time, test time and set
points: (a) automatic control display; (b) typical test plot;
(c) diagram of pneumatic test system.
(a)
Effect of Ambient Barometric
Pressure on Gage Pressure
Readings
(b)
Exhaust
Valves closed in 0.1 s
Pressure, kPa (lbf·in.–2)
If temperature remained constant and
uniform and no significant leakage
occurred during a pressure change leak
test period, the absolute pressure would
remain unchanged. Yet even when the
absolute pressure remains constant, the
gage pressure decreases as barometric
pressure increases, in accordance with
definitions of absolute and gage pressures
in this chapter. Conversely, if the
barometer rises during the test period, the
gage pressure would decrease by the same
pressure increment. These changes in
indicated gage pressure of the test volume
that result from variations in ambient
barometric pressure (and that are not
caused by leakage) are factored out of the
test data when a barometer or absolute
pressure gage is used to measure the
pressure used in computing the actual
leakage rate.
35 (5)
Stabilize
(wait)
0
Test pressure
decay
Fill
2s
4s
3s
1s
Time
(c)
Gage
Electronic Memory
Pressure Decay Leak
Testing System
Two types of pressure change leak testers
used for pressure decay leak testing are
the electronic memory type and the
differential pressure type. The electronic
memory type leak tester shown in Fig. 12
is widely used in pressure decay leak
testing. This system uses a (gage) pressure
Solenoid
valves
Pressure transducer
140 kPa (20 lbf·in.–2 gage)
Air supply
550 to 830 kPa
(80 to 120
lbf·in.–2 gage)
Test
part
Pressure
regulator
Atmosphere
Pressure
Quick
disconnect
Pressure Change and Flow Rate Techniques for Determining Leakage Rates
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
163
pressure leak testing in the range from 0
to 70 kPa (0 to 10 lbf·in.–2) gage, this
equipment is used with the reference
pressure port of the differential sensor
open to the ambient atmospheric
pressure. For pressure decay leak testing, a
reference chamber is used to equalize the
pressures acting on the differential
pressure sensor diaphragm at the start of
each leak test. This initial pressure value
can be stored in an electronic memory
before the pressure decay leak test.
Mass flow leak testing instruments are
used for applications where quantitative
flow measurements are required (Fig. 15).
Pressure decay instruments offer precision
of ± 0.02 Pa (3 × 10–6 lbf·in.–2).
chamber and the effects of adiabatic
heating during pressurization.
As an example, assume that leakage
rate tests are to be conducted on a test
system volume of 1.6 L (100 in.3)
pressurized to 140 kPa gage (20 lbf·in.–2
gage). The leakage rate sensitivity chart
(Fig. 13) for a 140 kPa (20 lbf·in.–2 gage)
transducer indicate that a pressure decay
period of about 20 s would be required to
achieve a leakage rate sensitivity of
5 × 10–3 Pa·m3·s–1 (5 × 10–4 std cm3·s–1).
See dashed lines A on Fig. 13.
Pressure Decay Leak
Testing
Figure 14 shows a variable capacitance
differential pressure test setup. For low
FIGURE 13. Graphical relationship between leakage rate sensitivity and test system volume for instrument shown in Fig. 12.
10
(100)
5
(50)
1
(10)
20 s
s
1
s
2
s
d
D
s
ec
ay
pe
60
rio
s
(10)
s
0.1
10
(5)
5
0.5
30
s
15
Leakage rate, 10–3 Pa·m3·s –1 (10–3 std cm3·s –1)
2s
0.05
0.01
(0.5)
B
A
(0.1)
0.01
0.02
(3.5×10–4) (7.1×10–4)
0.05
0.1
0.2
(1.8×10–3) (3.5×10–3) (7.1×10–3)
0.3
1
2
5
10
(0.011)
(0.035)
(0.071)
(0.18)
(0.35)
System volume, L (ft3)
Legend
A = 140 kPa (21 lbf·in.–2) transducer
B = 14 kPa (2 lbf·in.–2) transducer
164
Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
Absolute temperatures in degrees rankine
(°R) are derived form temperatures in
fahrenheit degrees by Eq. 3:
Ambient to Absolute Temperature
Measurement
Absolute zero temperature corresponds to
zero kelvin (0 K) and is equal to
–273.15 °C (–459.67 °F). Absolute kelvin
temperatures can be derived from
temperatures in other units by Eqs. 1, 2
or 4:
(3)
=
≅
459.67 + °F
460 + °F
Finally, absolute temperatures in degrees
kelvin (K) can be determined from
rankine temperature values by Eq. 4:
= 273.15 + °C
≅ 273 + °C
and from fahrenheit (°F) temperatures by
Eq. 2:
459.7 + °F
(2)
K =
1.8
460 + °F
≅
1.8
(1)
°R
K
(4)
K
=
°R
1.8
FIGURE 14. Schematic diagram of differential decay test setup.
Optional
pressure
transducer
Quick
Gage disconnect
Air
supply
Gage
Test
item
In
Out
Pressure
regulator
Quick
disconnect
Pressure
transducer
Quick
disconnect
Solenoid valves
In
High
Low
Out
Reference
chamber
Pressure
regulator
Optional pressure
switch
FIGURE 15. Schematic diagram of mass flow test setup.
Gages
Quick
disconnect
Solenoid
valve
Pressure
transducer
Air
supply
In
Quick
disconnect
Out
Test
item
Flow
Pressure
regulators
Out
In
Optional pressure
switch
Solenoid valves
Pressure
transducer
Pressure Change and Flow Rate Techniques for Determining Leakage Rates
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
165
Techniques for Surface
Thermometers in Pressure
Change Leakage Tests
Surface thermometers, such as those
shown in Fig. 16, may be used for small
volume systems during leak testing, where
it would be impractical to attempt to
measure the internal air temperature.
Temperature measurements must be made
during pressure change leak testing in any
case where temperature change can affect
the results of the pressure testing due to
the magnitude of the allowable pressure
change or the duration of the pressure
test. Surface thermometers must be held
tightly against the surface whose
temperature is to be measured. Any
suitable techniques such as tape, magnets,
couplant or clamps may be used to ensure
this firm and intimate contact between
the thermometer sensing surface and the
surface of the test object whose
temperature is to be measured. Procedures
and test reports for pressure hold tests
should specify the number and locations
of surface thermometers used during each
test.
The double nut in the center of surface
thermometers such as the types shown in
Fig. 16 should not be loosened or
tampered with by test operators or other
personnel because it is locked in position
to preserve the calibration setting of the
thermometer. Thermometers used for
testing should be calibrated periodically
by a qualified instrument laboratory to
provide assurance of their accuracy.
FIGURE 16. Surface thermometers on metal surfaces indicate
adjacent air temperature during leakage rate testing: (a) basic
surface thermometer; (b) surface thermometer with dual
permanent magnets in base for mounting on ferromagnetic
materials; (c) surface thermometer using both radiated and
conducted heat input.
(a)
Heated surface
(b)
Surface Thermometer Designs and
Mounting Techniques
Small, lightweight temperature indicating
surface thermometers are available in
various designs to cover several
temperature ranges from 0 to 300 °C (0 to
500 °F) or from 300 to 550 °C (550 to
1000 °F), for example. Typical accuracy is
±2 percent of full scale range. The basic
thermometer of Fig. 16a is designed for
horizontal or slightly curved surfaces. The
bimetallic coil in these instruments rests
directly on the surface whose temperature
is to be measured. The bimetallic spiral
coil of the sensor expands or contracts in
response to changes in temperature, thus
causing the dial itself to rotate. The
temperature of the surface is indicated by
the hooklike pointer outside the periphery
of the dial.
Figure 16b shows a type of surface
thermometer with three main parts: a
cover glass, a calibrated dial and indicator
and a magnetic base containing a
bimetallic thermal sensing element in an
inverted cup. The sensor is a bimetal alloy
designated by the applicable standard of
166
Leak Testing
Heated surface
(c)
Magnet
Heated surface
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
the American Society for Testing and
Materials and remains in permanent
calibration. In use, the sensing element
comes into virtual contact with the
surface whose temperature is to be
measured and provides a relatively fast
response, reaching full temperature
indications in about 3 min. The thermal
response time constant (time to achieve
one third of the temperature change)
varies from 0.06 s to 1.04 min depending
on the temperature range. The
thermometer is mounted by merely laying
it on any horizontal surface. On
ferromagnetic material surfaces, two
magnets in the base permit mounting in
any orientation. An ancillary, hand
adjusted pointer can be added to this
surface thermometer to remember specific
settings such as starting temperature or
final values.
Figure 16c shows a type of surface
thermometer that senses surface
temperature by conduction and radiation
effects. The base is applied to the surface
to be measured. Heat is transferred to the
base of the unit, which contains a
bimetallic element. Radiation from the
base inward causes the sensing element to
react, producing a resulting dial readout.
The bimetallic sensor is a specially
processed alloy that is preconditioned and
pretested for permanent calibration. The
instrument contains a highly reflective,
evaporated mirror that acts to protect the
sensor from the effects of external
radiation. This protective feature helps to
provide more accurate temperature
readings. The instrument is sealed against
entry of corrosive atmospheres. The
accuracy is ±2 percent of the full scale
range.
Dry Bulb Temperature
Measurements by
Resistance Thermometers
In pressure change leak tests of larger
structures, the temperature sensors in
general use are 100 Ω copper thermohm
detectors using a temperature sensitive
element of extremely pure copper wire,
wound into a helix and annealed to
minimize mechanical strain. This type of
construction provides a definite resistance
value for each temperature within the
range of the temperature detector. This
stability and accuracy ensures the
repeatability of measurements —
important in leakage rate calculations
because data to be analyzed are based
primarily on measuring changes in
temperatures and not on measuring the
actual temperature. Response time of the
copper wire temperature detectors for
90 percent of a temperature change is
about 40 s. The limit of error of the
detector is about ±0.03 °C (±0.05 °F) over
the temperature range from 0 to 120 °C
(32 to 250 °F).
Generally, suitable numbers of
resistance thermometers are located
throughout the volume of the structure
during leakage rate testing to provide an
adequate representation of internal
temperatures in each significant volume.
The number of detectors selected is a
function of the contained free air volume,
the configuration of the system under test
and the redundancy desired to ensure
representative contained air temperature
sampling if one or more temperature
sensors malfunctions. Each temperature
sensor is then assigned a volume fraction
based on the fraction of the total volume
under test. This volume fraction is a
temperature zone that may be determined
by prior temperature surveys and
represents the portion of the contained
gas or air that the individual sensor is
monitoring. The values of temperature
indicated for each temperature sensor are
recorded together with readings of
pressure sensors, at each interval during
the pressure change leakage test. These
temperature data are multiplied by the
fractional volumes they represent and the
weighted average contained air
temperature for the test volume is
computed and recorded for use in
correcting pressure indications for the
effects of temperature changes.
Sensors for Dew Point
Temperature
Measurements
The dew point temperature is a direct
indication of the amount of water vapor
present in the air contained within a test
volume subject to pressure change leakage
rate tests. If the temperature was reduced
to the dew point temperature, moisture
would condense on solid surfaces and
thus be temporarily removed from the
contained air. Vapor pressure due to
moisture evaporated into the contained
air adds to the total pressure measured by
most pressure detecting instruments used
in leakage rate testing. Two types of dew
point sensors used in leak testing are
aluminum oxide capacitance sensors and
resistance dependent sensors mounted on
thermoelectric cooling elements.
Capacitive Dew Point Sensors
Capacitance type dew point gages (also
known as aluminum oxide dew point
detectors) consist of a strip of metallic
aluminum anodized by a special process
to provide a porous oxide layer. A very
thin coating of gold is then evaporated
Pressure Change and Flow Rate Techniques for Determining Leakage Rates
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
167
over this oxide structure to provide a
conducting electrode. The aluminum base
metal and the gold layer electrode thus
form two electrodes with the dielectric
oxide layer between them, which serves as
an electrical capacitor. The concentration
of water vapor in the ambient air changes
the dielectric constant and so varies the
electrical capacitance of the sensing
element. Used as an impedance element
in an electronic circuit, this variable
capacitor produces output signals that
measure the dew point temperature in the
atmosphere contained within the system
under pressure change leakage test. The
system accuracy is typically ±1 °C
(±1.8 °F) over a dew/frost point
temperature range from –80 to +20 °C
(–110 to + 68 °F). The repeatability of
output signal readings is reportedly
±0.5 °C (±0.9 °F) in the dew point range
commonly encountered in leak testing.
Resistive Dew Point Sensors
Resistance dew point sensors are formed
on the surface of an insulating disk
consisting of epoxy filled fiberglass cloth.
A pair of intermeshing gold conductive
fingers provides electrodes for the surface
resistive element (the uncoated fiberglass
insulator). This surface resistance is
affected by moisture condensed on the
fiberglass insulator between the two
electrodes. The resistance sensing disk is
mounted on a two stage thermoelectric
cooler. Current supplied to the
thermoelectric cooler is controlled by
comparing the sensor resistance to that of
a fixed resistor. The dew point
determination is based on this surface
conductivity (which increases when liquid
water is formed by condensation on the
cooled surface of the sensor). The dew
point temperature range of this detector
system extends from –29 to + 57 °C
(–20 to +135 °F). The repeatability of
signals is in the range of ±0.3 C (±0.5 °F).
Correcting Pressure
Change Leak Test Data for
Changes in Vapor Pressure
The partial pressure of water vapor adds to
the true pressure of gases to produce the
total pressure of contained fluid measured
by the pressure sensors used in pressure
change leakage rate testing. If the partial
pressure of water vapor remained constant
throughout the duration of a leakage rate
test and constant throughout the test
volume, the value of this constant partial
pressure could be subtracted from the
total pressure measured to obtain the
pressure due to contained gases that
generally obey the ideal gas laws.
However, if the temperature changed,
168
Leak Testing
condensing water from the air or
evaporating more water into the air,
within a constant volume system under
test, the vapor pressure of the water
would change significantly. If no
correction were made for these nonideal
variations in vapor pressure, leakage
measurements by pressure change could
have considerable error. This error can be
avoided if all total pressure measurements
are corrected by subtraction of the known
water vapor pressure, so that leakage rate
calculations are based only on the
changes in the partial pressure of air (or
nitrogen or other pressurizing gas that
obeys the laws for ideal gases). Numerous
physical tables relate the partial pressure
of water vapor to dew point temperatures,
to temperatures of air in equilibrium over
water of the same temperature or to other
data such as relative humidity,
temperature and barometric pressure. For
example, the CRC Handbook of Chemistry
and Physics lists tables relating the
pressure of aqueous vapor over water
(torr) to temperature (°C).1 Steam tables
based on American Society of Mechanical
Engineers (ASME) data also relate the
partial pressure of water vapor (lbf·in.–2
absolute) to temperature (°F) as well as in
SI units of pressure (kPa) and
temperatures (°C and K).1
At the dew point temperature,
equilibrium exists between the partial
pressure of water vapor in air above a
surface on which water is condensing or
from which water is evaporating. Thus,
the dew point temperature measured
during leak testing can be related
immediately to the partial pressure of
water vapor at the location and
temperature of the dew point sensor
(Table 2).
Effect of Pressurization on
Dew Point Temperature
and Water Vapor Pressure
During pressurization of systems to be
tested for leakage by pressure change or
flow rate leakage tests, the partial pressure
of water vapor is increased in proportion
with the total pressure of contained air.
Thus, the dew point and probably the
relative humidity will also increase during
pressurization. Therefore, use of an air
dryer on the supply air during
pressurization is recommended. If a large
volume system (such as a nuclear power
reactor containment structure) is to be
tested and is provided with cooling coils
for the ventilation system, these cooling
systems should be used to minimize any
increases in dew point temperature during
pressurization. During the leakage rate
test, the dew point temperature should be
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
monitored for any changes in trends. A
sudden change in the rate of variation of
dew point temperature with time could
indicate water leakage.
TABLE 2. Water vapor pressures as a function of
dewpoint temperature in degree Celsius, in pascal and
in pound per square inch absolute.
Dewpoint Temperature
______________________
Vapor Pressure, Absolute
________________________
°C
(°F)
Pa
(lbf·in.–2)
–18
–17
–16
–15
–14
–13
–12
–11
–10
–9
–8
–7
–6
–5
–4
–3
–2
–1
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
(–0.4)
(1.4)
(3.2)
(5.0)
(6.8)
(8.6)
(10.4)
(12.2)
(14.0)
(15.8)
(17.6)
(19.4)
(21.2)
(23.0)
(24.8)
(26.6)
(28.4)
(30.2)
(32.0)
(33.8)
(35.6)
(37.4)
(39.2)
(41.0)
(42.8)
(44.6)
(46.4)
(48.2)
(50.0)
(51.8)
(53.6)
(55.4)
(57.2)
(59.0)
(60.8)
(62.6)
(64.4)
(66.2)
(68.0)
(69.8)
(71.6)
(73.4)
(75.2)
(77.0)
(78.8)
(80.6)
(82.4)
(84.2)
(86.0)
(87.8)
(89.6)
(91.4)
(93.2)
(95.0)
(96.8)
(98.6)
(100.4)
124.8
137.2
152.4
166.2
181.3
197.9
216.5
235.8
257.9
281.3
307.5
335.8
370.3
402.0
436.4
473.7
514.4
558.5
610.2
657.1
706.0
757.7
813.6
872.2
934.9
1001.8
1072.8
1148.0
1228.0
1312.8
1402.4
1497.6
1598.2
1705.1
1818.2
1937.4
2063.6
2196.7
2337.3
2486.3
2643.5
2808.3
2982.7
3166.8
3360.5
3564.6
3779.0
4004.5
4241.7
4491.3
4754.0
5029.1
5318.0
5621.3
5939.9
6273.6
6623.8
0.0181
0.0199
0.0221
0.0241
0.0263
0.0287
0.0314
0.0342
0.0374
0.0408
0.0446
0.0487
0.0537
0.0583
0.0633
0.0687
0.0746
0.0810
0.0885
0.0953
0.1024
0.1099
0.1180
0.1265
0.1356
0.1453
0.1556
0.1665
0.1781
0.1904
0.2034
0.2172
0.2318
0.2473
0.2637
0.2810
0.2993
0.3186
0.3390
0.3606
0.3834
0.4073
0.4326
0.4593
0.4874
0.5170
0.5481
0.5808
0.6152
0.6514
0.6895
0.7294
0.7713
0.8153
0.8615
0.9099
0.9607
Determining Gas Pressure
from Total Pressure and
Water Vapor Pressure
The air, nitrogen or other typical
pressurizing gas used in pressure change
leakage tests is selected so that it obeys
the ideal gas laws relating pressure,
temperature and volume. The water vapor
contained in the pressurizing gas fails to
obey these ideal gas laws, yet it
contributes a partial pressure which adds
to the ideal gas pressure to equal the total
gas pressure measured by pressure sensing
instruments during the leakage tests. To
permit valid estimations of true gas
leakage rates, the partial pressure Pv of
water vapor must be subtracted from the
total absolute pressure P to obtain the true
gas pressure Pg as shown in Eq. 5 for net
ideal gas pressure:
(5)
=
Pg
P
−
Pv
Equation 5 applies to absolute pressure
only, in any single system of pressure
units. Water vapor pressure varies in air as
a function of dew point temperature, in SI
units (see also Table 2).
Calculation of Leakage Rate
by Pressure Change Test
(Constant Temperature)
If the test is of short duration and it is
known that temperature has not changed
during a pressure hold test (or if
temperature conditions remain constant),
the test requires only measurement of gage
pressure. In this case, the time rate pressure
change can be calculated from Eq. 6:
(6)
∆P
∆t
=
P1 − P2
∆t
As an example of a calculation using
Eq. 6, suppose that a pressure hold test is
conducted on a system with an allowable
pressure loss rate of 7 kPa (1 lbf·in.–2) in
30 min. If the initial gage pressure was
400 kPa (56.0 lbf·in.–2) at time 13:00 and
the final gage pressure was 396 kPa
(55.4 lbf·in.–2) at time 13:30, Eq. 6 indicates
that the time rate of pressure loss is
∆P
∆t
=
=
=
P1 − P2
∆t
4 kPa
30 min
=
=
400 − 396
30
130 Pa ⋅ min –1
2.2 Pa ⋅ s –1
Pressure Change and Flow Rate Techniques for Determining Leakage Rates
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
169
In English units, the same test calculation
would appear as:
∆P
∆t
P1 − P2
=
=
∆t
56.0 − 55.4
30
0.6 lbf ⋅ in.−2
=
30 min
=
1.2 lbf ⋅ in. –2 ⋅ h −1
=
3.33
×
10 –3 lbf ⋅ in.−2 ⋅ s −1
The measured rate of pressure loss is less
than the allowable pressure loss rate of
7 kPa (1 lbf·in.–2) in 30 min, indicating
that the system under test is acceptable
because its leakage rate is below the
specified maximum allowable leakage
rate.
Calculation of Leakage
Rate by Pressure Change
Test (Constant Volume)
During a pressure change leakage test of a
system with fixed volume, the initial
volume V1 and the final volume V2
remain essentially identical. Thus, for the
special case of constant volume systems
under test, V1 = V2 and Eq. 7 applies to
the pressure change leak test period:
(7)
from variations in ambient barometric
pressure (and which are not caused by
leakage) are factored out of the test data
when a barometer or absolute pressure
gage is used to measure the absolute
pressure.
P1
P2
=
T1
T2
Correcting Pressure
Change Leak Test Data for
Changes in Temperature
When a short duration pressure hold test
is conducted under varying temperature
conditions and requires measurement of
both gage pressure and temperature but
does not require measurement of
barometric pressure, the barometric
pressure is assumed to be one standard
atmosphere (101.3 kPa or 14.7 lbf·in.–2).
The pressure loss per unit of time is then
determined from the initial gage pressure
P1 and temperature T1 and the final gage
pressure P2 and the final temperature T2,
by means of Eq. 8. The temperatures must
be absolute temperatures and the absolute
pressures may be taken as the gage
pressures plus an assumed standard
barometric pressure.
For gage pressures in kilopascal and
temperatures in degree celsius, using SI
units and measuring time in seconds, the
pressure change rate is given by Eq. 8:
(8)
∆P
∆t
=
−
or
P1
=
T1
P2
T2
As can be seen from the first form of
Eq. 7, absolute pressure varies in direct
proportion with absolute temperature. In
the absence of significant leakage, the
absolute pressure increases in proportion
with an increase in contained absolute gas
temperature. Conversely, lowering the gas
temperature lowers the absolute internal
gas pressure proportionately.
÷
If temperatures remained constant and
uniform and no significant leakage
occurred during a pressure change leak
test period, the absolute pressure would
remain unchanged. This is in contrast to
the gage pressure, which increases as
barometric pressure decreases by the same
pressure increment when no significant
leakage occurs.
These changes in indicated gage
pressure of the test volume which result
170
Leak Testing
)
(P
(T
2
2
)(
+ 273)}
)]
+ 101 T1 + 273
÷ ∆t
For gage pressures in pound per square
inch and temperatures in degree
fahrenheit, using English units and
measuring time in minutes:
(9)
∆P
∆t
=
−
÷
Effect of Ambient Barometric
Pressure on Absolute Pressure
Gage Readings
[(
 P + 101
 1

[(
 P + 14.7
 1

(P
(T
2
2
)
)(
+ 460 )} ÷
+ 14.7 T1 + 460
)]
∆t
For absolute pressures in kilopascal and
temperatures in degree celsius, using SI
units and measuring time in second:
(10)
∆P
∆t
P1 − P2
=
T1 + 273
T2 + 273
∆t
For absolute pressures in pound per
square inch and temperatures in degree
fahrenheit, using English units and
measuring time in minute:
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
(11)
∆P
∆t
=

T1 + 460 
 P1 − P2 T + 460 


2
∆T
For absolute pressures and absolute
temperatures, the correction takes on the
simpler form of Eq. 12:
(12)
∆P
∆t
=

T1 
 P1 − P2

T2 

∆T
where all terms are expressed in SI units
or where all terms are expressed in
English units.
Determining Mass of
Contained Gas for Pressure
Change Leakage Tests of
Large Volume Systems
The time rates of leakage are determined
by the changes in the total mass of air,
nitrogen or other ideal pressurizing gas
contained within the test volume V, after
corrections for temperature T and water
vapor pressure Pv. In the absolute test
technique, the ideal gas law can be
expressed in the form of Eq. 13, for the
case in which the test volume remains
constant:
(13)
W
=
K1
V
R
P ′ − Pv
T
where W is measured mass of contained
(ideal) gas or air, kilogram (or pound); V is
internal free volume of system under test,
cubic meter (or cubic foot), constant; R is
individual gas constant. (For air, R = 287
J·kg–1·K–1 or 53.35 ft-lbf ·lbm–1·°R–1) P is
total absolute pressure in test volume,
pascal (or lbf·in.–2 absolute); Pv is partial
pressure of water vapor in contained air,
pascal (or lbf·in.–2 absolute); T is mean
absolute temperature of air contained in
test volume kelvin (or degree rankine); K1
is 1 (for SI units). K1 = 144 (for English
units for conversion from pressure in
lbf·in.–2 to lbf·ft–2). Typically, the leakage
rate can be determined from the change
in contained air mass through a
succession of test point data readings or
by subtracting the final mass (at the end
of a test period) from the initial contained
mass (at the beginning of the test period).
The mass change must be divided by the
time interval between successive readings
or between initial and final readings, to
provide the time rate of leakage. The mass
leakage rate would then be given by
Eq. 14:
(14) Q t
=
∆W
∆t
where ∆W is Wstart – Wend = change in
contained mass during test interval; ∆T is
tend – tstart = time interval between start
and end of test interval.
Determining Mass Loss of
Contained Gas for Pressure
Decay Tests of Large
Volume Systems
When the test volume is constant, the
mass of contained air or gas at the
beginning of the test period is given by
Eq. 15:
(15) W1
=
P1
V
R T1
The mass of contained air at the end of
the test period is given by Eq. 16:
(16) W2
=
P2
V
RT2
The mass loss due to leakage during the
test period is then given by Eq. 17:
(17) W1 − W2
=
 P1
 T
 1
–
P2  V
T2  R
In Eq. 15 through 17, W1 is initial mass of
contained air (kilogram); W2 is final mass
of contained air (kilogram); P1 is initial
absolute test pressure (pascal); P2 is final
absolute test pressure (pascal); T1 is initial
contained air temperature, kelvin
(= °C + 273.15); T2 is final contained air
temperature (kelvin); V is test volume
(cubic meter); and R is gas constant for air
(287 J·kg–1·K–1).
Determining Leakage Rate
in Volume Units at
Standard Temperature and
Pressure
The standard conditions for volume loss
leakage rates are as follows: Ps is standard
pressure, 101.325 kPa (14.696 lbf·in.–2
absolute); Ts is standard temperature,
20 °C or 293.15 K (68 °F or 527.67 °R); Vs
is volume of air at standard conditions
corresponding to a particular mass W. The
mass of air at standard conditions is
related to the standard volume Vs by
Eq. 18:
Pressure Change and Flow Rate Techniques for Determining Leakage Rates
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
171
(18) Ws
=
When actual pressure change leakage rate
test data are used, the leakage rate Q s in SI
units is given by Eq. 23:
Ps Vs
R Ts
The volume of air at standard conditions
is given in terms of mass W by Eq. 19:
(19)
Vs
=
(23) Q s
W R Ts
Ps
The leakage rate Q s in standard volume
units is given by Eq. 20:
(20) Q s
=
Vs1 − Vs 2
∆t
W1 RTs
=
=
−
Ps
W2 R Ts
Ps
∆t
R Ts
Ps ∆ t
(W
1
− W2
)
When the actual pressure change leakage
test data are used (as measured at test
temperatures T1 and T2 and with
corresponding test pressures P1 and P2),
the standard leakage rate Q s in standard
volume units is given by Eq. 21:
(21) Q s
=
V
∆t
Ts
Ps
 P1
P2 
T − T 
 1
2 
When SI units are used in Eq. 20 or 21,
test volume V is given in cubic meter; the
time interval ∆t, in second; the pressure P,
in pascal; and the temperature T, in
kelvin (K = °C + 273.15). Ps is simply
dropped; the leakage rate is then given in
pascal cubic meter per second.
When English units are used, the test
volume is measured in cubic foot; the
time, in hour; the pressure, in pound per
square inch; and the temperature, degree
rankine (= °F + 459.7). The leakage rate Qs
is then given in standard cubic foot per
hour.
Determining Leakage Rate
in SI Units at Standard
Temperature and Pressure
It should be noted that the leakage rate Q
in SI units has been expressed in this
book in units of Pa·m3·s–1, which is the
product of volume and pressure, divided
by time. In this case, the leakage rate Q s
in SI units is given by Eq. 22:
(22) Q s
172
Leak Testing
=
RTs
t
(W1
− W2 )
=
 P
V
P2 
Ts  1 −

t
T2 
 T1
Ps
where Q is leakage rate (Pa·m3·s–1); t is test
duration (second); R is individual gas
constant, J·kg–1·K–1 (for air, R = 287
J·kg–1·K–1); V is test volume (cubic meter);
Ts is standard absolute temperature, K
(i.e., 293 K); W1 is mass of contained air
or gas at beginning of test (kilogram); P1 is
pressure at beginning of test (pascal); W2
is mass of contained air or gas at end of
test (kilogram); P2 is pressure at end of test
(pascal); T1 is absolute temperature at
beginning of test, kelvin (K = 273 + °C1);
T2 is absolute temperature at end of test,
kelvin (K = 273 + °C2); the subscript s
denotes standard. Ps is standard pressure
of 101.3 kPa.
Determining Mass of
Contained Air after
Correction of Water Vapor
Content
The actual pressure of ideal pressurizing
gas (air, nitrogen or other gases obeying
the ideal gas law) can be determined by
subtracting the pressure of contained
water vapor from the total pressure, in
accordance with Eq. 5. The mass Wg of
ideal gas is given in terms of total pressure
P minus the pressure of water vapor Pv:
(24) Wg
=
V
RT
(P
− Pv
)
Equation 24 would apply to quantities
expressed in SI units as listed for Eq. 17.
In practical (mixed) units (used in shop
or field leak tests in industry before 1981),
Eq. 25 gives the mass of contained air (or
ideal gas) after correction for water vapor
content:
(25) Wg
=
144
V
RT
(P
− Pv )
where Wg is mass of contained air
(pound); V is internal free volume of
containment (cubic foot); R is gas
constant for air, 53.35 ft-lbf·ft-lbm–1·°R–1; T
is mean absolute temperature (dry bulb)
of contained air (degree rankine); P is
total absolute pressure in containment
(lbf·in.–2 absolute); and Pv is partial
pressure of water vapor in containment
(lbf·in.–2 absolute).
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
Leakage Rate Test Data
Obtained by Absolute Test
Technique in English Units
The analysis used with absolute pressure
leakage rate tests consists of determining
the mass of air in the containment, using
the ideal gas law, at each time point
during the test and using a straight line
least squares analysis to estimate the
leakage rate. Errors in the determined
masses are assumed to be equally variable
(i.e., the slope and intercept of the line
are estimated by ordinary as opposed to
weighted least squares) and uncorrelated.
An upper one-sided confidence limit for
the leakage rate is based on normal
regression theory (i.e., the masses are
related by a straight line and deviations
from that line are normally distributed)
and a technique due to Fieller for finding
confidence limits for ratios of means of
normally distributed random variables.
For each time point ti, the corresponding
mass of contained air Wi is determined
directly from the application of the ideal
gas law as given in Eq. 26:
(26) Wi
144 V
R
=
Pi − Pv i
×
Ti
A linear least squares fit of the data is
then made according to the relation
(Wi)a = Ati + B.
The estimate of the leakage rate is a
function of both the slope and the
intercept of the regression line (percent
per day):
(27) Q am
=
− 2400
A
B
In Eq. 27, the term A represents the
slope of the least squares straight line. The
term B indicates the intercept of this
straight line with a vertical line drawn
through the time scale point for t = 0. The
numerical constant 2400 is the product of
the number of hours in a day (24) and the
multiplier (100) for a percentage
calculated from a ratio. The negative sign
(–) indicates that, for a pressure decay test,
the regression line slopes downward from
the initial point at ti = 0 to later points at
ti = tn.
Effects of Time Duration of
Pressure Change Leakage
Rate Tests
For short duration absolute pressure
change leak tests, such as a 2 h pressure
hold test, the change in atmospheric
pressure is usually insignificant and
standard barometric pressure can be
assumed to exist. (Care is needed to avoid
this assumption during passage of a cold
front or low pressure storm system,
because rapid changes in barometric
pressure can accompany such storm
periods.) If the allowable pressure loss per
unit of time is large enough, it may also
be possible to eliminate measurement of
temperature and to measure only pressure
and time. For very short duration pressure
tests (such as 15 to 120 min), the leak
testing procedure may require only
measurement of gage pressure and time
(in constant volume systems).
For longer duration tests such as 24 h
pressure hold or leakage rate tests, it is
most likely that specifications for test
procedures will require measurement of
both temperature and barometric pressure
(or absolute pressure) because of the larger
atmospheric changes that could occur in
these two test variables. It may also be
necessary to measure dew point
temperature to account for variations in
water vapor pressure with temperature.
Analysis Techniques for
Pressure Change Leakage
Test Data2,3
Three techniques for analysis of data
obtained during pressure change leakage
rate testing of pressurized test systems are
(1) the mass point analysis technique,
(2) the leakage rate point analysis
technique based on total time from start
of test and (3) the leakage rate point
analysis technique based on test interval
data.
Mass Point Technique of Analysis
of Pressure Change Leak Test Data
In mass point data analysis, data from an
absolute technique leak testing system are
reduced to a value for the mass W of air
within a pressurized test volume, by
application of the ideal gas law. The test
data consist of a time sequence of
independent values for the contained air
mass. Figure 17a is a graphical illustration
of a short sequence of mass point test
data, plotted vertically as a function of
elapsed test time, shown horizontally. The
successive sets of test data are identified
by subscripts n = 0, 1, 2, 3, 4 … k. The
term Wn is the value of the air mass inside
the test volume at the time tn. In practice,
Wn often is represented in percentages of
the initial air mass at the start of the
leakage test at time t = 0 and the elapsed
test time is often recorded in hours.
(Later, the leakage rates may be stated in
percentages of initial mass change per
Pressure Change and Flow Rate Techniques for Determining Leakage Rates
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
173
Figure 17 illustrates a simple example of
the leakage rate point-to-point technique
of analysis of leakage rate test data.
Individual leakage rates Qn are calculated
from the mass differences between
successive adjacent test points:
(28) Q n
=
Wn − Wn −1
t n − t n −1
The values of this point-to-point
leakage rate are equivalent to the slopes of
the lines labeled as Q 1, Q 2 … Q 5 in
Fig. 17b. If these point-to-point leakage
rates are then plotted on a new graph as a
function of elapsed test time, the result is
similar to Fig. 17c. Here, the computed
leakage rate in percentage change per day
of the initial contained mass W0 is shown
on the vertical scale and the elapsed test
time on the horizontal scale. Positive
values for leakage are shown in Fig. 17c
when the slope of the line Q n in Fig. 17b
is downward.
Negative values of leakage are shown in
Fig. 17c when the slope of the
corresponding line in Fig. 17b is upward.
The sloping line in Fig. 17c indicates the
leakage rate trend with elapsed time.
When this trend line flattens, it indicates
establishment of the leakage rate with
additional test time serving only to
increase the reliability of the data. When
test data are taken at regular time
intervals, there is no implicit weighting of
data. The effective leakage rate is simply
the arithmetic mean of all the individual
leakage rates when these data are taken at
roughly equal time intervals. This greatly
simplifies online data analysis during
pressure change leakage rate testing.
FIGURE 17. Various statistical techniques for analyzing
leakage rate from identical point-by-point test data during
pressure change leakage rate test, after Fleshood2 and Lau3:
(a) leak testing data with computed air mass plotted as
function of elapsed test time tn for the mass point analysis
technique, where slope of dashed line from W0 to W5
indicates overall leakage rate Q = (W0 – W5)/(t5 – t0);
(b) leakage rates calculated from mass differences between
adjacent test points, where slopes of short lines indicate
incremental leakage rates, Qn = (Wn –1 – Wn)/(tn – tn–1), valid
for n greater than zero; (c) leakage rate trend line calculated
by linear least squares analysis of incremental leakage rates
Qn shown in Fig. 17b.
(a)
Mass W (relative units)
Point-to-Point Analysis of Pressure
Change Leakage Rate Test Data
initial mass W0 at the start of the test and
the mass Wn for the most recent data
point, as slopes of individual lines,
Qn = (Wn – W0)/(tn – t0). Each successive
leakage calculation is therefore based on a
longer period of time, tn – t0. A different
leakage rate may thus be computed for
W0
W1
W3
W2
W5
W4
t0
t1
t2
t3
t4
t5
Time during test (h)
(b)
Leakage rates
Mass W (relative units)
day.) In Fig. 17a where k = 5, t 0 is the time
when leak testing begins (zero hours) and
W0 is the mass of air within the test
volume when leak testing begins. W5 is
the mass of contained air after an elapsed
test time of t5 in hours.
Q1
Q3
Q5
Q2
Q4
t0
t1
t2
t3
t4
t5
Time during test (h)
Various statistical techniques may be used
for analyzing leakage rates from identical
point-by-point leak test data during
pressure change leakage rate test, where
slopes of lines equal leakage rates.2,3
Figure 18a illustrates the total time
technique of calculating leakage rates
based on the mass difference between the
174
Leak Testing
(c)
Leakage rate (percent per day)
Leakage Rate Total Time
Technique of Analysis of
Pressure Change Leak Test
Data2,3
+
Q2
Q4
Q1
Linear
least squares fit
0
Q3
Q5
–
t0
t1
t2
t3
t4
t5
Time during test (h)
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
Wn − W0
t n − t0
With this analysis technique, it is not
proper to assume that the effective
leakage rate for the total test period is a
simple average of all the individual
leakage rates calculated by Eq. 29. Each
successive leakage rate is calculated from
the contained air mass change over a
longer elapsed time period. The result of
averaging Qn leakage rates is heavily
weighted toward the larger values of n
(longer total times, tn – t0).
Caution must be taken to time weight
each datum appropriately. Also, the
instrumentation errors for small n values
will show up as relatively large deviations
in the analysis. See also Fig. 18d for linear
least squares fit evaluation of total time
leakage test data.
Figures 18a shows the test volume air
mass vertically, as a function of elapsed
test time shown horizontally, by the
individual test point dots. The several
lines connecting the initial W0 mass point
(at upper left) to the successive Wn mass
points at different elapsed test times have
slopes corresponding to the leakage rates
computed for each of the successively
longer elapsed testing times. Figure 18b
shows the leakage rates plotted vertically
in percentages of W0 change, ±(Wn–0/W0)
per day, as a function of elapsed testing
time (shown horizontally) from start of
test to most recent mass measurement, in
percent of initial mass change per day. In
this case, all values of leakage rate shown
in Fig. 18b are positive, because all line
slopes in Fig. 18a are downward.
(a)
Q1
Mass W
(relative units)
=
Q3
Q2
Q5
Q4
t0
t1
t2
t3
t4
t5
Time during test (h)
(b)
Linear least squares fit
using a sloping line
Leakage rate
(percent per day)
(29) Q n
FIGURE 18. Statistical techniques for pressure change leakage
rate test: (a) leakage rates calculated from mass difference
between starting mass and mass at test time; (b) leakage
rates plotted as function of elapsed test time in percent of
initial mass change per day; (c) average leakage rate;
(d) total time test data with least squares fit to eliminate
time dependency; (e) linear least squares fit drawn through
mass point leakage test data shown in Fig. 17a.
t0
t1
t2
t3
t4
t5
Time during test (h)
(c)
Leakage rate
(percent per day)
each test point following the initial point,
by Eq. 29 for total time leakage rate:
Q1
t1
Q5
Q4
Q2
t0
Linear least squares fit
using a constant
Q3
t2
t3
t4
t5
Time during test (h)
(30) Q n′′
=
W0 − Wn
t n − t0
Q3
Q2
t0
t1
Q4
t2
t3
t4
t5
Time during test (h)
(e)
W1
W0
Linear least squares fit
using a sloping line
W3
W2
W5
W4
t0
This leakage rate corresponds to that
indicated as Q 5 in Fig. 18a. Had the
arbitrary test period been different in
Linear least squares fit
using a constant
Q5
Q1
Leakage rate
(percent per day)
From the example illustrated by Fig. 17, it
is self evident that numerous different
values for leakage rate could be derived
from the same initial test data from an
absolute technique test. For example, in
many leak tests, it is considered
appropriate to determine the leakage rates
simply from the initial mass or pressure
within an enclosure and the final mass or
pressure at the end of some arbitrary
testing time. In this case, Eq. 30 gives the
endpoint leakage rate:
(d)
Mass W
(relative units)
Limitations of Time Dependent
Test Data from Pressure Change
Leak Tests
t1
t2
t3
t4
t5
Time during test (h)
Pressure Change and Flow Rate Techniques for Determining Leakage Rates
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
175
length, the leakage rate might have been
equally well determined as Q1, Q2 or any
other Qn shown on a graph similar to that
of Fig. 18a. For this reason, techniques of
statistical analysis are often used to lend
credence to leak testing data, where
measured leakage rates are significantly
influenced by test conditions, some of
which may be chosen arbitrarily. Several
different statistical techniques are
described next, to illustrate their
possibilities.
Estimating Constant Leakage
Rates from Average or Least
Squares Fit of Data
It might be reasonable to assume that
leakage rates are essentially constant
throughout the pressure change tests
when the absolute pressure is essentially
constant throughout the test and the size
of any leakage path should not change.
For this case, it would be necessary to fit a
constant to the test data, as shown by the
horizontal lines of Fig. 18c for the
point-to-point analysis (same as average
leakage rate) and of Fig. 18d to eliminate
time dependency in total time analysis of
the leak testing data of Fig. 17. The least
squares relationship requires that the sum
of the squares of the deviation (Qn – Q)
should be a minimum. This is equivalent
to requiring that the derivative (d·dQ–1) of
this sum of mean squares with respect to
Q should be equal to zero, as shown in
Eq. 31 for the condition for minimum:
(31)
0
2
d 
 Q − Q1 +
dQ 
(
=
+ …+
(Q −
)
(Q − Q )
2
2
2
Qk 

)
Therefore, in the case in which
point-to-point leakage rates are taken at
roughly equal time intervals during the
pressure change leak testing period, the
linear least squares fit is equal to the
simple arithmetic means value of all of
the individual values for leakage rates.
This greatly simplifies online data analysis
during the leakage test, where the best
linear least squares fit to the test data can
be computed continuously during testing
operations by the simple average of
Eq. 32:
(32) Q a
=
Q1 + Q 2 + … + Q k
k
Note that in Eq. 32, k is the total number
of leakage measurements made at equally
space time intervals, after t = 0.
176
Leak Testing
Applying Least Squares Fit to a
Line of Mass Point Leak Test Data
Figure 17a shows an example illustrating
data collected in the mass point analysis
technique before any data analysis has
taken place. In Fig. 18e, these data have
been fitted to a sloping line by a least
squares technique. The slope of this line is
drawn through mass point leakage test
data shown in Fig. 17a. Leakage rate,
percent per day = 100 [(W0 – Wn)/W0]
[24/(tn – t0)], where t is time (hour).
If it can be assumed that the leakage
rate is constant with respect to elapsed
time during the leakage test, the data are
appropriate for analysis by the technique
of least squares because of the
independent nature of this type of
analysis, an error during testing will result
in only one bad datum and will not
materially affect the leak testing results. If
mass point analysis and fitting by least
mean squares is carried out continuously
as each set of data is taken during the
mass point analysis, results are consistent,
although not identical. When two hourly
sets of data are combined to make a third
set, the results always average as expected.
With techniques using real time data
analysis and graphical plotting in real
time during tests, the approach to
uniform rates of leakage can be seen and
tests extended or terminated as
appropriate to the quality and consistency
of data.
The mass point of 95 percent
confidence ranges from 0.05 to 0.2 times
the measured leakage rate. By comparison,
the 95 percent confidence interval may
range from one half to twice the
measured leakage rate with the total time
technique and from two to 20 times the
measured leakage rate with the pointto-point technique. For these reasons,
many organizations prefer the mass point
technique with continuous data analysis
to the alternative techniques of analysis.
Formulas for Computing Least
Squares Line Fitting Mass Point
Leak Test Data
The theoretical basis for using least
squares techniques to compute a leakage
rate lies in the so-called Gauss-Markoff
theorem. As applied to the measurement
of leakage rates, the theorem states that, if
the linear relationship between W and t is
appropriate and if the W values are
independent and equally variable, the best
estimators of the slope and intercept of the
line are given by least squares analysis.
Here, best means two things: (1) the
estimators are not biased and (2) the
estimators have the smallest variances of
any other unbiased estimators that might
be derived from arbitrary linear
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
LT.05 LAYOUT 11/8/04 2:16 PM Page 177
combinations of the W values. The least
squares line is given by Eq. 33:
n= K
∑ (W
=
(33) Wa
At
+
B
where the slope A and intercept B are
given, respectively, by Eq. 34 and 35:
(34)
(35)
A
n
=
B =
∑t W
n ∑t
i
∑W ∑ t
n ∑t
i
∑W ∑ t
− (∑ t )
−
i
2
i
2
i
2
i
i
=
− 2400
σ
n =1
=
n
i
2
i
∑t W ∑t
− (∑ t )
−
i
i
−
2
where S is the standard deviation, Wi is
the computed mass at time ti (from
Eqs. 33 or 35) and n is the number of the
leak test measurements. Now, let the
quantity K be defined by Eq. 38:
S
=
(38) K
n
i
∑t
−
2
i
(∑ t )
2
i
2
i
Each ti is the elapsed time between the
clock time at which the initial reading is
taken and the clock time at which the ith
reading is taken. Thus, t1 = 0 in all test
situations, t2 is the length of the time
elapsed before the next reading and so on.
In most test situations, the time intervals
between tests will be constant but the
formulas for A and B do not require
constancy.
The leakage rate is expressed as the
ratio of the rate of change of mass to the
mass in the containment at time t1 = 0.
Because values of ti have units of hours
and percentage daily leakage rates are
desired, the mass point leakage rate is
expressed as a positive number of Eq. 36:
(36) Q am
=
(37) S
− Wi )
2
1
A
B
Note that B — not the mass W0 measured
at the initial time — is used as the
denominator of Q am. B is the better
measure of the contained mass because
W0 has the same error structure as the Wi
values.
The uncertainty in the estimated value
Q am is assessed in terms of the standard
deviations of A and B and their
covariance, followed by the computation
of an upper limit of the 95 percent
confidence level for Q am. In what follows,
the full details are spelled out. Conditions
are stated that result in considerable
simplification applicable to most leakage
test situations.
Formulas for Computing Standard
Deviation in Mass Point Leak Test
Data
The estimate of the common standard
deviation (following from the equally
variable assumption) of the masses with
respect to the line is given by Eq. 37:
Then, the standard deviation of the slope
is given by Eq. 39:
(39)
=
SA
K
n
The standard deviation of the intercept is
given by Eq. 40:
(40)
=
SB
K
∑t
2
i
And the covariance of the slope and
intercept is given by Eq. 41:
(41)
SAB
=
( ∑t )
K2 −
i
Confidence Limits for Mass
Point Leak Test Data2,3
The confidence limit is a measure of the
statistical consistency in test data.
Figures 19 and 20 illustrate the meaning
of the confidence limit in terms of the
normal Gaussian distribution of data with
random errors. The shaded area of the
curve in Fig. 19 is equal to 95 percent of
the area within the total Gaussian
distribution curve when the latter is
integrated from –X to +X. A 95 percent
confidence limit means that 95 percent of
the measurements will fall within the
shaded range of leakage rates. It can also
mean that, if another identical test was
run, then statistically there is a 95 percent
chance that the calculated leakage rate
will be within the shaded range of Fig. 19.
In Fig. 20 the confidence limit is
plotted vertically as a function of the
dispersion index (plotted horizontally).
The units of the dispersion index scale are
the standard deviation of Eq. 37. The 95
percent confidence limit corresponds to
the dispersion index value equal to three
standard deviations. With a dispersion
index equal to only one standard
Pressure Change and Flow Rate Techniques for Determining Leakage Rates
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
177
FIGURE 19. Normal Gaussian distribution curve. Shaded area includes 95 percent of the measurements in a
normal distribution. After Fleshood2 and Lau.3
Cumulative percents
99.5 percent
Leakage rate (relative units)
97.5 percent
95 percent
2.5 percent
0.5 percent
0.5 percent
0
10
20
30
40
50
60
70
80
90
100
Percent of measurements
FIGURE 20. Percentage confidence limit plotted as a function
of dispersion index, measured in standard deviation units (σ).
Ordinate or curve shows what confidence level applies for
each value of dispersion index shown on horizontal scale.
After Fleshood2 and Lau.3
deviation, the confidence limit taken
from the curve of Fig. 20 would be
reduced to about 70 percent. The
universal standard deviation σ is defined
by Eq. 42:
n= k
100
Confidence limit for normal distribution (percent)
90
(42) S
80
70
σ
=
n
)2
n =1
k
Because the standard deviation and the
confidence limit can be calculated easily
with the aid of programmable hand
calculators, microprocessors or
minicomputers as the leak test progresses,
the test operator can readily determine
what percentage confidence level is
attained. Decisions can then be made as
to whether the test should be extended to
attain the required degree of statistical
confidence or discontinue until repairs are
made to the test system or unreliable
instrumentation is replaced.
60
50
40
30
20
0
0.5
1
1.5
2
2.5
Dispersion index, standard deviation unit (σ)
178
=
∑ (Q − Q
Leak Testing
3
Formulas for Calculating
Approximate and Exact
Limits of Confidence Level
The data of Table 3 relate the 95th
percentile t0.95 of the test data distribution
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
to selected values for the number of
degrees of freedom, dF = n – 2, where n is
equal to the number of leak test
measurements of the mass W of
contained air at the corresponding
elapsed test time t, following the initial
measurement at time t = 0.
The standard deviation of the slope SA
was defined by Eq. 39 and the standard
deviation of the intercept SB was defined
by Eq. 40 for the least squares line defined
by Eq. 33, namely W = At + B. In most
leakage testing situations, the ratio SB·B–1
is very small compared with the ratio
SA·A–1. Thus, an approximate upper limit
(UCL) for the 95 percent confidence level
of the percentage leakage rate of Eq. 36 is
given by Eq. 43:
SA
B
3
(43) ~ UCL = Q am + 2.4 × 10 t 0.95
Values for t0.95 are selected from the data
of Table 3, with dF = (n – 2).
For the case of n = 20 or more test
points, following the initial data at time
t = 0, the values of t0.95 can be determined
from Eq. 44:
(44) t 0.95
=
1.645
−
2.4
(
1.576
+
n−2
)
2
+
n−2
57.6
(n − 2)
3
where dF = n – 2.
The adequacy of the approximate
confidence level computed by Eq. 43 is
measured in terms of its closeness to the
exact Fieller type limit derived from the
assumption that the Wi values are
normally distributed about the straight
line.4 Experience with Type A leak tests
has shown this approximation to be
entirely adequate. However, to obtain the
exact upper confidence limit, let:
(45) a
=
B2
−
t 02.95 SB2
(46) b
=
AB
−
t 0.95 SAB
c
=
A
2
−
t 0.95 SA
(47)
2
2
2
The exact upper one sided limit of a 95
percent confidence level for the
percentage per day leakage rate is given by
Eq. 48:
(48) Q am : UCL
=
×
10 3
− 2.4
×
b −
b2 − a c
a
Possible Reasons for
Rejection of Erroneous
Data from Pressure
Change Leak Test
To obtain adequate accuracy in pressure
change leakage rate testing, the
instruments used for leakage
measurements must be very accurate and
sensitive. Nevertheless, fluctuations of
leak test data points cannot be avoided.
An outlying observation or an outlier is a
datum widely different from the
remaining observations in the data set.
The outlying observation may be the
result of an error in calculating a
numerical value and could probably be
corrected if properly identified. An outlier
could also result from an instrument error
or from an error in reading the
instrument’s indication. If this is known
to be the case, the false reading should be
removed from the data set. Hence, the
testing engineer is always confronted with
TABLE 3. Tabular relationship between
number of sets of leak testing data
following initial W0 and t0, and 95th
percentile of distribution, t 0.95, as a
function of number dF of degrees of
freedom, after Fleshood.2
n
dF
t0.95
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
34
∞
6.314
2.920
2.353
2.132
2.015
1.943
1.895
1.860
1.833
1.812
1.796
1.782
1.771
1.761
1.753
1.746
1.740
1.734
1.729
1.725
1.721
1.717
1.714
1.711
1.708
1.706
1.703
1.701
1.699
1.697
1.684
1.671
1.658
1.645
Pressure Change and Flow Rate Techniques for Determining Leakage Rates
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
179
the task of determining when a test
datum is spurious or bad. A bad datum
must be rejected; otherwise it will increase
the standard deviation unduly. However,
an apparently bad datum cannot be
eliminated arbitrarily. There must be a
valid basis for such rejection, such as a
valid statistical criterion that identifies a
true outlying observation.
To use the technique proposed by
Tietjen in pressure change tests, let ti
denote the ith time (hour) for the ith
reading, Wi the corresponding air mass
and (Wi)a = Ati – B the corresponding
predicted mass from Eq. 33. Then the ith
residual wi = Wi – (Wi)a has a standard si:
(t − t )
∑ (t − t )
2
Responsible Usage of
Criteria for Leakage Test
Data Rejection
Where a statistical criterion for testing an
outlying datum is permissible, it cannot be
applied selectively. That is, one should not
apply the criterion to an outlier if its
inclusion in the calculations would reduce
the calculated leakage rate unless one is
also prepared to reject an outlier whose
inclusion would increase the calculated
leakage rate or its upper confidence limit.
For this reason, it is appropriate for the
user to determine in advance of the leakage
test whether or not the criterion is to be
used and what the rejection level for data
points will be if this criterion is applied.
It should be noted also that, if a high
percentage of test points (such as two or
more in 20 points) has to be rejected, the
test engineer must conclude that either
(1) the test instrumentation and procedure
used on the leak test must be improved or
(2) some systematic errors are not
accounted for. In either case, the deviation
does not follow a normal Gaussian
distribution to which the statistical
criterion could be properly applied. On
the other hand, if most of the leak testing
data are widely scattered, then an
additional widely scattered datum is likely
to be found to be acceptable according to
the statistical criterion for identification of
a true outlier. In this case, the standard
deviations in measured leakage rates will
be larger and the confidence limit will be
smaller than the typical 95 percent upper
confidence limit desired.
(49) si
1
= S 1 −
i
−
n
a
2
i
a
In Eq. 49, the term S (standard
deviation estimated from least square line)
is given by Eq. 50:
(50)
S2 =
=
or
(51)
S
=
∑
1
n − 2
∑W
i
( ) 
W − W
i
i

2
a
2
n − 2
∑W
1
n − 2
i
2
and t is given by Eq. 52:
(52) t a
=
1
n
∑t
i
The standardized residual ri = wi·si–1 is
next computed and the potential outlier
Wi is the observation whose absolute
value of the standardized residual ri is the
largest. Once D = max |ri| is located, it is
compared to a value in Table 4, to
determine whether this quantity is
significant. If D exceeds the table value,
Wi is declared an outlier.
For a leakage rate test in which the
data are collected at equal time intervals,
Eq. 49 reduces to Eq. 53:
(i − i )
∑ (i −i )
2
(53)
si
1
= S 1 −
−
n
a
2
a
in which ia is defined by Eq. 54:
Data Rejection Criterion
for Regression Data from
Pressure Change Leak Test
Most traditional tests for an outlying
observation are not appropriate for testing
for an outlier in a regression situation,
such as pressure change leakage rate
testing, because the standard error of
residual varies with time. An acceptable
test criterion for a single outlier in a
simple linear regression, however, is given
by Tietjen et al.5
180
Leak Testing
(54) i a
=
=
=
1
n
∑i
1+ 2 + 3+ … + n
n + 1
n
2
A still simpler form is shown in Eq. 55:
(55)
si
= S 1 −
1
n
−
(
12 i − i a
(
)(
)
2
)
n n +1 n −1
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
TABLE 4. Values of critical deviation ratio
D for data rejection for a one-sided
statistical test, used in criterion for
outlier data in containment leakage test.
Number of
Observations
(n)
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
5 Percent
Rejection
Level (D)
1 Percent
Rejection
Level
1.41
1.71
1.92
2.07
2.19
2.28
2.35
2.43
2.48
2.52
2.57
2.61
2.64
2.68
2.71
2.74
2.76
2.79
2.82
2.84
2.85
2.89
2.90
2.92
2.93
2.95
2.96
2.97
2.99
3.00
3.01
3.02
3.03
3.04
3.06
3.07
3.08
3.09
3.09
3.10
3.11
3.12
3.13
3.14
3.15
3.15
3.16
3.17
3.17
3.18
3.18
3.19
3.19
3.20
3.21
3.21
3.21
1.41
1.73
1.97
2.16
2.31
2.43
2.53
2.64
2.70
2.76
2.80
2.87
2.92
2.96
2.99
3.03
3.06
3.09
3.12
3.15
3.17
3.19
3.21
3.23
3.25
3.26
3.28
3.29
3.31
3.32
3.33
3.34
3.35
3.36
3.37
3.38
3.39
3.40
3.40
3.41
3.42
3.42
3.43
3.44
3.45
3.45
3.46
3.46
3.47
3.47
3.48
3.48
3.49
3.49
3.50
3.50
3.50
Example Application of
Criterion Technique for
Outlier Datum
This example illustrates the technique
used for identifying and evaluating an
outlier datum. For every point in time,
Table 5 shows the containment air mass,
its deviation from the linear least squares
fit, the standard error of the residual and
the standardized residual. In this example,
with data generated at 15 min intervals
from an actual test, the number of data
points n = 36. With the measurements
made at equal time intervals using Eq. 54:
36 + 1
=
ia
=
2
18.5
and
∑W
i
=
2
28 848.83
and using Eq. 50,
=
S
28 848.83
36 − 2
The estimated standard deviation of
the containment air mass from the linear
least square fit is given by Eq. 55 so that
si
=
×
28 848.83
36
−
1 −
1
36
2
−
(
)
(36) (37) (35)
12 i − 18.5
2
The maximum absolute standardized
residual is found from the last column of
Table 5 for i = 28, a where D = |ri| = 2.08.
The absolute magnitude is indicated
by |ri|.)
From Table 4, it is seen that, for n = 36,
a D statistic as large as 2.08 occurs more
often that 5 percent of the time; hence,
the potential outlier should not be
rejected on statistical ground. Because the
largest standardized deviation is not
rejected, no other datum can be rejected
statistically, either.
If for the datum, the residual were –96
instead of –59.07, just to illustrate the
5 percent data rejection criterion, one
would have obtained D = 3.38. From
Table 5, it is seen that the datum would
have occurred less than 5 percent of the
time and could have been rejected
statistically.
Pressure Change and Flow Rate Techniques for Determining Leakage Rates
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
181
System Ability to Measure
Leakage Rate
The purpose of a leak testing
instrumentation selection guide is to
determine the ability of a specific
instrumentation system to measure the
overall leakage rate of a pressurized
system adequately. This selection guide is
not based on a statistical analysis of the
leakage rate calculations, but has been
developed for the purpose of selection of
instrumentation adequate for the required
leakage measurements. In evaluations
made using one guide,6 the errors of
individual instruments used for
measurement of pressure and temperature
or dew point are combined using a
statistical root-sum-square formula:
(56) δQ = ISG = ±
2.4
×
t
2
×
10 3
2
 eP 
e 
e 
2 P  + 2 T  + 2 v 
 P 
T 
 P 
2
ISG is the instrumentation selection
guide; δQ is the standard deviation δ of
the leakage rate Q (percent per day); t is
the test duration (hour); P is the
containment atmosphere total absolute
pressure; Pv is the containment
atmosphere partial pressure of water
vapor; T is the containment atmosphere
weighted average absolute dry bulb
temperature; e is the error associated with
measurement of change in a given
parameter; E is the error associated with
sensor sensitivity.
TABLE 5. Example of calculations for a single outlier test datum in pressure change test
for leakage rate.
Datum
i
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
182
Leak Testing
Air Mass
Wi
Linear
Least Squares Fit
W
Residual from
Least Squares Fit
wi = Wi – (Wi)a
Standard
Error of Residual
si
Standardized
Residual
ri = wi·si
D = |ri|
735 478.1
735 473.5
735 475.8
735 451.1
735 439.8
735 449.6
735 444.2
735 426.6
735 415.1
735 396.7
735 391.3
735 426.3
735 440.7
735 424.8
735 432.3
735 435.3
735 409.1
735 423.5
735 436.4
735 436.4
735 391.8
735 392.1
735 452.8
735 455.5
735 448.9
735 371.3
735 387.9
735 359.6
735 395.4
735 375.0
735 407.8
735 445.5
735 446.5
735 447.0
735 464.2
735 437.0
735 443.37
735 442.46
735 441.54
735 440.63
735 439.71
735 438.80
735 437.88
735 436.97
735 436.06
735 435.14
735 434.22
735 433.31
735 432.39
735 431.48
735 430.56
735 429.65
735 428.73
735 427.82
735 426.90
735 425.99
735 425.07
735 424.16
735 423.24
735 422.33
735 421.41
735 420.50
735 419.58
735 418.67
735 417.75
735 416.84
735 415.92
735 415.01
735 414.09
735 413.18
735 412.26
735 411.35
34.73
31.04
34.26
10.47
0.09
10.80
6.32
–10.37
–20.95
–38.44
–42.92
–7.01
8.31
–6.68
1.74
5.65
–19.63
–4.32
9.50
10.41
–33.27
–32.06
29.56
33.17
27.49
–49.20
–31.68
–59.07
–22.35
–41.84
–8.12
30.49
32.41
33.82
51.94
25.65
27.53
27.67
27.79
27.91
28.02
28.12
28.21
28.30
28.38
38.45
28.51
28.56
28.61
28.64
28.68
28.70
28.71
28.72
28.72
28.71
28.70
28.68
28.64
28.61
28.56
28.51
28.45
28.38
28.30
28.21
28.12
28.02
27.91
27.79
27.67
27.53
1.26
1.12
1.23
0.38
0.00
0.38
0.22
–0.37
–0.74
–1.35
–1.51
–0.25
0.29
–0.23
0.06
0.20
–0.68
0.15
0.33
0.36
–1.16
–1.12
1.03
1.16
0.96
–1.73
–1.11
–2.08
–0.79
–1.48
–0.29
1.09
1.16
1.22
1.88
0.93
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
Nature of Systematic
Errors and Random Errors
In estimating the magnitude of the
uncertainty or error in the value assigned
to a quantity (mass of air in containment)
as the result of measurements, a
distinction must be made between two
general classes of error, systematic and
random. Systematic errors are those errors
associated with a difference between the
true value and the measured parameter
produced by predictable or identifiable
effects. Calibration of the leakage rate
measuring system traceable to the
National Institute of Standards and
Technology removes systematic errors or
reduces them to an acceptable magnitude.
Random errors are those whose
magnitude and sign fluctuate in a manner
that cannot be predicted from a
knowledge of the measurement system,
the system calibration certification or the
conditions of measurement.
Techniques for Verification
of Accuracy in Leakage
Test Measurements
An acceptable technique to verify that a
significant calibration shift or system
change has not occurred is to make a
definite, known change in the magnitude
of the measured value using a different,
independent, calibrated instrument. This
is accomplished with the verification test.
Such comparison provides a check to
verify that a significant calibration shift or
other system change has not occurred and
that the measurement system systematic
error has remained essentially constant.
Therefore, a successful verification test
confirms that the leakage rate test system
systematic error is within acceptable
limits. Any other error associated with
leakage rate measurement is then due to
random error.
For verifying the validity of the leakage
rate test measurements during the change
leak tests, the following supplemental
techniques described in Appendix C of
ANSI/ANS-56.8-1981 may be used.6
calibrated flow meter or rotameter. The
leak orifice is selected to provide a flow
under the test pressure condition
equivalent to 75 to 125 percent of the
leakage rate specified for the acceptance
test.
The test procedure involves placing the
calibrated leak system into operation after
the leakage rate test in progress is
completed. The flow meter readings are
then recorded at least hourly.
Concurrently, readings of the
containment system leakage measuring
system record the composite leakage of
both the containment system leakage rate
and the superimposed leakage rate.
The readings of the flow meter as a
function of time enable calculation of the
average leakage rate through the
calibrated orifice. From the analysis of the
readings taken with the leakage measuring
system, the composite leakage rate Qc is
determined.
The duration of the superimposed
leakage verification test depends on the
leakage rate involved and generally
requires at least 4 h with a minimum of
ten sets of data.
Supplemental Technique Using
Metered Mass Change
A mass step change verification test using
a metered quantity of air. A small
quantity of air is either metered into or
out of the containment over a short time
interval. This mass change indicated by
the leak test instrumentation prior to and
following the metered mass change is
compared to the metered mass change.
The mass step change verification test
is conducted as follows. At the end of the
leak test a mass of air is metered through
a flow meter, either into or out of the
containment over a short time interval.
This metered mass change is compared to
the mass change indicated by the leak test
instrumentation before and after the
metered mass change. The change in mass
calculated from the test instrumentation
must agree within 25 percent with the
metered mass change.
Supplemental Technique Using
Calibrated Leak6
A calibrated or measurable leak is
intentionally superimposed on the
existing leaks in a system under test. A
practical and simple arrangement uses the
orifice leak of a microadjustable
instrument flow valve installed at a
convenient penetration of the
containment system. The flow through
the valve is measured by means of a
Pressure Change and Flow Rate Techniques for Determining Leakage Rates
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
183
PART 2. Pressure Change Leakage Rate Tests in
Pressurized Systems
Operating Principles of
Pressure Change Leakage
Rate Testing
Leakage rate testing by measurement of
pressure changes in closed volumes
requires that the system under test be
maintained at a pressure other than
ambient atmospheric pressure. Pressure
change leak tests can be made with either
an evacuated or a pressurized test system.
The leakage rate Q is equal to the
measured pressure change ∆P multiplied
by the test system’s internal volume V
and divided by the time interval ∆t,
required for the change in systems
pressure to occur:
(57) Q
= V
∆P
∆T
where Q is leakage rate (Pa·m3·s–1); V is
enclosed system volume (cubic meter);
∆P = P1 – P2, which is pressure change
during leak test (pascal); ∆t = t2 – t1, which
is time interval during leak test (second).
The pressure change leak testing
procedure is used primarily for leakage
measurement in large systems. However,
with minor modifications, the pressure
change technique can be used to measure
leakage rates on test systems of any size.
This procedure is used only for
measurement of leakage and is not well
suited for location of individual leaks.
However, a leak may be localized to a
closed part of a system under test by
pressure change test techniques.
Sensitivity of Pressurized
Mode Leakage Tests by
Pressure Change
Techniques
The sensitivity of leakage measurement
during leak testing of pressurized systems
with the pressure change technique
depends on the minimum detectable
magnitude of pressure variation. Static
pressure is measured at the start, at
intervals and at the end of the leak testing
period. The sensitivity of this static
leakage measurement largely depends on
the time duration of the test and the
sensitivity and accuracy of the pressure
184
Leak Testing
measuring instruments. In the absence of
uncontrolled temperature changes or
severe outgassing effects, longer time
intervals between initial and final
measurements permit more sensitive
measurements of leakage rate.
The accuracy of measurement of
leakage rates in the pressurized mode of
pressure loss leak testing depends on how
precisely the test volume V is calculated
and on how accurately the changes in
pressure and temperature can be
measured. If the leakage rate is measured
as a percentage of total enclosed fluid
(mass) lost per unit of time, then
precision in calculating the enclosed
volume may not be required. When using
properly calibrated pressure measuring
instruments in the pressurized mode, the
accuracy of leakage measurement by the
pressure loss technique can often be
traced to the National Institute of
Standards and Technology.
Sources of Error in
Pressurized Mode Leakage
Tests by Pressure Change
Techniques
The test procedure for the pressurized
mode of leakage measurement consists of
filling the test system with gas and
observing any pressure decrease. The
fundamental relationship is given in
Eq. 57.
Two large sources of error exist in this
technique. The volume of the test system
is difficult to calculate for a large or
complex system; however, it can be
measured by the additional leakage
technique, which is also known as a
verification test or a proof test in practice.
An additional known leak is added to the
system under test. The system volume is
then calculated from the effect of the
additional leakage on the observed rate of
pressure decrease.
The second source of error inherent in
the pressure change technique exists
when temperature variations during the
test cycle tend to vary the pressure in the
system. This error can be corrected by
measuring system temperature during the
leak test. The pressure effect of
temperature variations can be calculated
by using the ideal gas laws. In an
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
alternative technique for correction for
interfering effects, a reference volume is
placed in the system under test and the
variations of pressure differential between
this closed reference system and the test
system are observed. Specific illustrative
examples of such calculations are given
later in this chapter.
Advantages and
Limitations of Pressure
Change Techniques of
Leak Testing
Two major advantages of the pressure
change technique of leak testing are the
following.
1. Instrumented large scale pressure or
vacuum systems can often be leak
tested by using pressure gages already
installed on the system to be tested.
2. No special tracer gas is required.
Two major disadvantages of the
pressure change technique of leak testing
are the following.
1. The time required for leak testing can
be rather long.
2. This test technique does not permit
precise leak location without auxiliary
techniques.
Pressure change leak tests can be
conducted on any contained volume that
will withstand the internal pressure used
to apply the necessary pressure differential
across the boundaries of the test volume.
Pressure Change Leakage
Rate Testing of Constant
or Variable Volume
Systems
Pressure change leak testing is a
nondestructive test technique used for
determining the total leakage rate through
the walls or pressure boundaries of a
structure tested at a specific pressure.
Pressure change leak tests can be
conducted on any contained volume that
will withstand either an internal pressure
differential (pressure system) or an
external pressure differential (vacuum
system) across the boundary of the test
volume. For constant volume or variable
volume pressure systems with gage
pressure greater than atmospheric
pressure, the pressure change leak test is
also commonly identified by names such
as pressure hold test, pressure loss test,
pressure decay test or leakage rate test.
A constant volume system is a rigid
structure such as a pressure vessel where
the physical change in the size of the
system due to temperature variation is so
small relative to total contained volume
that it can be ignored. A variable volume
system is a flexible structure such as a
vapor tank in which the volume changes
to maintain a uniform internal pressure.
For large volume systems, the gas
temperature and dewpoint in the system
under test should be measured if possible
throughout the time period used for the
pressure change leak test.
Selection of Pressurizing
Gases for Pressure Change
Tests for Leakage Rates
Pressurizing gases used for pressure
change leakage rate testing should obey
the ideal gas laws to a reasonable degree.
The most commonly acceptable gases in
this category are air, nitrogen, helium,
argon and carbon dioxide. Use should
never be made of hazardous pressurizing
gases such as toxic gases or oxygen (which
supports combustion of oils, grease or
hydrocarbons). Similarly, combustible
gases such as propane, butane or
acetylene should never be used for
pressurizing because of the dangers of
explosion.
The common halogen rich tracer gases
(such as refrigerant-12 or refrigerant-22)
should not be used as pressurizing gases
for absolute pressure leak testing because
they do not obey the ideal gas law and
can produce erroneous leak testing results.
If refrigerant gases have been used in a
system as the tracer gas for preliminary
halogen detector probe leak testing, these
chlorinated hydrocarbons must be purged
from the system under test prior to
performing a pressure change test for
leakage rates.
Precautions in Preparation
for Pressure Change
Leakage Rate Testing
The following preliminary leak testing
techniques and practices are desirable
before pressure change leak testing during
fabrication or erection of large items such
as pressure vessels or liners, test channels,
double gasket flange interspaces or
airlocks, for example. Before conducting a
pressure change test, preliminary leak
testing should be performed to detect and
eliminate leakage from connections
external to the test object. Otherwise,
such external leaks could affect the results
of the pressure change leak test. The type
of preliminary testing that should be
performed is usually given in the written
Pressure Change and Flow Rate Techniques for Determining Leakage Rates
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
185
procedure for leak testing of the specific
products or assemblies.
When preliminary leak testing includes
a halogen detector probe test, the halogen
mixture should be purged from the test
system before conducting a pressure
change test. Also, before starting the
pressure change test, the operator should
always close the inlet isolation valve and
disconnect the pressurizing line or
manifold; then, tests to locate all leaks
should be performed on this valve
connection and the pressure gage
connection.
If adverse working conditions are
encountered during the day work shift, it
is often best to perform a short duration
pressure hold leak test of a small volume
system during a less busy shift or when
there is less interference. A longer
overnight leak testing period with more
stable ambient temperature conditions
may make it possible to pass a test object
or a channel test zone which otherwise
might improperly have appeared to have
failed during the usual 1 or 2 h leak test
during variable daytime conditions. For
such reasons, if a test object is on the
borderline of acceptance for a pressure
hold leak test, it is advisable to continue
the test overnight or during some other
convenient longer period not subject to
interference from other work activities.
Typical Test Sequence for
Pressure Change Leak
Testing in Industry
After completing all required preliminary
testing and after purging of the test
system (if halogen rich refrigerant was
previously used), the pressure change
leakage rate test is performed in the
following steps:
1. A calibrated pressure gage is connected
to the contained volume under test.
When necessary, calibrated equipment
to measure dry bulb temperature and
dewpoint temperature (humidity) is
also installed and verified after
installation.
2. A pressurizing line is then attached to
a valve connection on the test system.
The test object is pressurized to the
designated test pressure (usually with
compressed air). The pressurized test
system is next isolated from the
pressurizing source with the valving
system. The pressurizing source is then
disconnected and a solution film
bubble emission test is next performed
on the seat and stem of the
pressurizing connection valve.
3. The pressure gage is observed to detect
any consistent loss in pressure not
related to temperature change. If the
186
Leak Testing
pressure remains reasonably stable, the
leak test can be started. If the pressure
constantly decreases more rapidly
than the allowable rate of pressure
decrease, additional preliminary
testing for leakage should be
performed.
4. Only after it has been established that
no detrimental leakage exists in
external connections, valves or other
components should the pressure
change leak test be started and test
data be recorded.
5. If, during the course of a pressure
change leakage rate test, any leak
testing instruments malfunction or
become damaged, they should be
replaced with properly functioning
instruments (if these instruments are
indispensable to the satisfactory
completion of the test). Then the
leakage rate test should be repeated
from the start.
6. A pressure change leakage rate test
may be concluded at the end of the
required test period if the magnitude
of the pressure loss or leakage is
within the specified allowable rate. If
the test results are borderline,
consideration should be given to
continuing the test time period to
increase the reliability of the test data.
If the pressure loss or leakage rate is in
excess of the allowable limits, the
system should be reinspected by other
testing techniques to detect the
location of the excess leakage.
7. When leaks with unacceptable leakage
rates are located, each such leak
should be repaired; then local retests
should be used to prove that the
leakage has been eliminated or
reduced to an acceptable level for each
leak. Finally, the entire system should
be retested by the specified pressure
change leak testing technique to
ensure that total leakage rates are
within acceptable limits.
Relation of Pressure to
Temperature (Volume
Constant)
Calculations of leakage rates from
absolute pressure readings in constant
volume test systems depend on test
variables including test time, temperature
and pressure. For tests of large systems, it
is also necessary to consider the effects of
water vapor pressure within the contained
volume. The static relation between the
pressure, volume and temperature of a
fixed of gas can be written as Eq. 58:
(58)
PV
T
=
constant
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
where P is absolute pressure (pascal or
lbf·in.–2 absolute); V is volume of
container (cubic meter or cubic inch); T is
absolute temperature (kelvin or degree
rankine). Consistent units, such as SI only
or English units only, should be used for
each term in Eq. 58 and in succeeding
equations relating the same parameters.
The basic equation for pressure change
leak testing used when comparing two
different conditions for a given mass of
the same gas (derived from Eq. 58 for test
conditions 1 and 2) is given by Eq. 59:
(59)
P1 V1
T1
P2 V2
T2
=
or
P1
P2
T1
T2
=
V2
V1
P1
P2
=
T1
T2
=
P2
∆P
=
P1
−
P2
T1
T2
To calculate the pressure change per unit
of time, use can be made of Eq. 62, in
which the time duration of the test
(between successive readings in a
sequence of readings or between start and
finish of a leak test) is taken as ∆t:
∆P
∆t
=
P1
−
P2
∆t
T1
T2
Extending the time duration or length of
a pressure change leakage test will
increase the magnitude of the pressure
change and usually result in an increase
in the accuracy and reliability of the leak
test results.
Calculation of Pressure
Change with Gage
Pressure and Thermometer
Readings
or
P1
(61)
(62)
For a pressure change of a given
(constant volume) system, the initial
volume V1 and final volume V2 remain
essentially the same. Therefore, for the
constant volume test systems, V1 = V2 and
Eq. 59 can be written more simply:
(60)
the total system leakage rate for the case
of the specific test pressure selected for
the leak test. Working equations for these
calculations are presented below.
If the pressure change during the test is
designated by ∆P, Eq. 61 corrects for a
change in temperature.
T1
T2
As can be seen from the first form of
Eq. 60, absolute pressure varies in direct
proportion with the absolute temperature.
In the absence of significant leakage, the
absolute pressure increases in proportion
with an increase in contained gas
temperature. Conversely, lowering the gas
temperature lowers the absolute internal
gas pressure proportionately.
Calculation of Pressure
Change with Absolute
Pressure and Temperature
Readings
The pressure change leak test is performed
by pressurizing a closed system to a
specific pressure and isolating the system.
Time, temperature (internal) and system
pressure are recorded systematically for
some test period. For large volume
systems, dewpoint would also be
measured to permit determination of the
partial pressure of water vapor.
Comparison of initial pressure P1 and
final pressure P2 can be used to determine
With small systems, pressures are
sometimes measured as gage pressures and
gas temperatures are measured with
ordinary thermometers or surface
temperature indicators on the celsius or
fahrenheit temperature scales. These
pressures must be converted to absolute
pressures and the temperatures must be
corrected to absolute temperatures in
kelvin or degree rankine. If the pressure
change test is made under conditions that
do not require measurement of the
barometric pressure, the barometric
pressure can be assumed to be one
standard atmosphere (101.3 kPa or
14.7 lbf·in.–2 absolute). The gas pressure
change is computed by either Eq. 8 for
celsius temperatures or Eq. 9 for
fahrenheit temperatures.
If barometric readings of the pressure
of the earth’s atmosphere are required and
barometric pressures vary, each individual
gage pressure measurement must first be
corrected to the absolute pressure value by
Eq. 63:
(63)
P
=
Pgage
+
P barometer
Pressure Change and Flow Rate Techniques for Determining Leakage Rates
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
187
where P is absolute pressure (kilopascal or
lbf·in.–2 absolute); Pgage is gage pressure
(kilopascal or lbf·in.–2); Pbarometer is
barometric pressure (kilopascal or lbf·in.–2)
obtained in uncorrected form from local
weather bureau or read from a precision
barometer and converted to pressure
units. Where pressures are measured in
other units such as torr, inch of mercury
or foot of water, the pressures must be
converted consistently either to English
units or preferably to SI units. After
conversions have been made, the rate of
absolute pressure change can again be
calculated by use of Eqs. 10 or 11.
test. In addition, the system displays the
ambient pressure and temperature
conditions. Flow measurements, vital to
leakage verification tests, are also integral
functions. This system accommodates the
superimposed leak tests technique or the
pumpback technique (the mass change
verification leak test). The leakage test
data are represented by a visual display, a
printed record of the raw test data and a
concurrent minicomputer calculation of
the leakage rates, in several forms.
Data Acquisition, Analysis
and Recording Systems for
Leakage Rate Testing
Figure 21 shows a schematic diagram of
the components and system used in an
integrated leak testing system. Typical
leakage rate computations are based on
measurements of the changes in the
absolute pressure, water vapor pressure
and the dry bulb temperature. The
absolute pressure is measured with a fused
quartz Bourdon tube. The low internal
viscosity of fused quartz makes it the most
perfectly elastic material available. This
type of pressure sensor has no measurable
hysteresis. It also has fast response, high
resolution and high accuracy. The water
vapor pressure is measured by use of
chilled mirror dewpoint sensors and is
presented to the minicomputer as a
dewpoint temperature in degrees celsius
or fahrenheit.
The dry bulb temperature is measured
by resistance temperature detectors and is
also presented to the minicomputer as
digital data. Because the changes in the
test parameters are small in magnitude, all
input sensors must be capable of high
sensitivity, accuracy, repeatability and
resolution. Similar high accuracy, high
resolution and reliability are required of
the electronic networks and digital
computer analyses.
Data recording for large scale pressure
change leakage tests is made simpler by
sophisticated numerical data acquisition
systems. There systems automatically
multiplex the conditioned signals from
the pressure, temperature, dewpoint and
flow measuring sensors (during the
verification phase of leak testing) at preset
automatically timed intervals. Data are
transmitted through an interface for
numerical analysis by computer, recorded
on magnetic tape or disk systems,
displayed by printout or graphical
recordings and evaluated by error analysis
and statistical techniques.
In many cases, this numerical test data
analysis system can analyze the data by
progressive analysis (with least mean
squares fit to straight line approximations
of leakage as a function of testing time).
Computers provide the fastest and most
accurate technique for analysis of the
pressure change leak test data. The data
can be fed into the computer directly
from the acquisition system interface,
from tape or manually from printer or
recorder readouts. This absolute technique
analysis of leakage rate may be performed
by mass point or leakage rate point-topoint, point-to-point cumulative and total
time statistical analysis techniques.
Minicomputer Integrated
Leakage Rate
Measurement System
Integrated leakage rate measurement
systems are available that include all
components from input sensors to
minicomputer analyses of test data. This
system will measure and record the
absolute pressure, the dewpoint
temperature and the dry bulb temperature
of the air within the system under leakage
188
Leak Testing
Components of Integrated
Leakage Rate Measurement
System
Microprocessor Data Acquisition
and Analyses with Leakage Rate
Measurement System
A microprocessor (minicomputer)
controls the minicomputer data
acquisition and raw data recording
system. The microprocessor system
includes both read-only memory (ROM)
and random access memory (RAM), a
scanner system and various interfaces
with sensor and output system
components. Digital data for pressure, dry
bulb temperature and dewpoint
temperature are presented in ASCII
(American Standard Code for Information
Interchange) to the computer which then
operates on these raw test data and
calculates the leakage rate (see discussion
below).
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
∆P
∆T
Example of Analysis of
Data from Pressure Hold
Test of Small Volume Test
Object
Table 6 shows leak test data analysis for
pressure hold tests of a small volume test
object with an allowable temperature
corrected pressure loss of 0.5 lbf·in.–2 in
2 h. In these tests, corrections for
variations in barometric pressure were not
required.
Analysis of data for the first pressure
hold test on day one by means of Eq. 8
shows that the test object has failed the
requirements of the pressure hold test:
=
[(42.0 + 14.7)
− (42.0 + 14.7 )
89.5 + 460 
–1
⋅2 h
95.8 + 460 
= 56.7 − 56.7
549.5
loss in 2 h
555.8
= 0.6 lbf ⋅ in. –2 loss in 2 h
= 0.3 lbf ⋅ in. –2 ⋅ h –1
This leakage rate exceeds the maximum
acceptable temperature corrected pressure
loss of 0.5 lbf·in.–2 in 2 h, so the test
results are not satisfactory. A leak was
located in the pressure zone and the welds
repaired. The pressure hold test was then
repeated four days later with the results
shown in the last two columns of the
table in the left hand column. In this
case, analysis of the pressure hold test
data showed:
FIGURE 21. Information flow diagram for minicomputer controlled integrated leakage rate measurement system using a
microcomputer, dual disk memory and instrument display console.
Containment
Console
Digital
Containment
pressure
Quartz manometers
Verification air flow
Turbine flow meters
Preset up/down
counter
Analog
Microprocessor data
conversion
Digital
RS-232-C
Input/output
port
Resistance
temperature
detectors
Resistance
temperature
detectors (signal
conditioning)
Analog
Digital
Analog
Scanner
Dew point
hygrometers
Hygrometer
control conditioning
circuitry
Analog
Digital data encoder
Digital
Manual
scanner
Linear variable
differential and other
transducers
Analog
Structural
integrity test
Panel meter
Analog
Pressure Change and Flow Rate Techniques for Determining Leakage Rates
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
189
∆P
=
∆T
56.7 − 56.7
2h
550.0
550.7 + 56.7 − 56.6
2h
= 0.1 lbf ⋅ in. –2 loss in 2 h
= 0.05 lbf ⋅ in. –2 ⋅ h –1
This leakage rate is well below the
maximum allowable temperature
corrected pressure loss rate of 0.5 lbf·in.–2
in 2 h, so the test object is now
acceptable.
The calculations in SI would reflect the
fact that 1 lbf·in.–2 ≅ 6.9 kPa.
Determining If Pressure
Hold Test is Completed or
Should Be Extended
A pressure hold leakage rate test may be
concluded at the end of the required test
period if the magnitude of the pressure
loss is within (lower than) the specified
allowable rate of pressure loss. If the test
results are subject to question, the test can
be continued over a longer time period to
increase the reliability of the test data. If
the extended test confirms that the actual
pressure loss rate is in excess of the
specified allowable limits, the system
should be reinspected to detect the
locations of the excess leakage. The
system leaks should then be repaired and
the system retested to the same
specifications and procedures.
Leak Testing Techniques
Using Cyclic
Repressurization with
Compressor
Intermittent operation of a compressor
can be used for leakage measurements by
evaluating the load cycle (duty cycle)
when a compressor of known capacity
maintains a specified pressure in the
system under test. (The duty cycle is the
ratio of the time the pump operates, on
time, to the total time testing — on time
plus off time for the pump.) With large
rates of leakage, the compressor must
operate for a large proportion of the test
time. With low leakage rates, the same
compressor need operate only
occasionally for relatively short time
periods. If it is desired to measure the
absolute value of leakage by this
technique, the capacity of the compressor
at the test operation pressure should be
known or be determined in a separate
compressor calibration test.
When leak testing by the cyclic
pressurization technique, operators need
two stopwatches and a rapid response
pressure gage with a clear scale. Prior to
the first measurements, the compressor is
allowed to charge the system under test to
its normal operating (or delivery)
pressure. The compressor control valve is
then shut off to isolate the test vessel. The
compressor is allowed to operate under no
load conditions while the pressure in the
test vessel falls off to a suitable lower
pressure limit value, well below the initial
compressor delivery pressure. When this
pressure limit is reached, one stopwatch is
started and the compressor is put under
load by manual opening of the isolation
control valve. During this operation step,
the compressor pumps air or gas into the
test vessel to raise its internal pressure.
This first stopwatch is stopped when the
test pressure has risen to a predetermined
upper pressure limit.
When this upper limit is reached, the
compressor is cut off by closing the
isolation valve of the test vessel and the
second stopwatch is started. When the
internal pressure of the test vessel or
system again falls to the original lower
pressure limit, the second stopwatch is
stopped and reset. Then the first watch is
started as the compressor is again put
under load to repressurize the test system.
A new cycle of pressurization is initiated
and the alternating stopwatch readings for
pressurization time and leakage time are
taken. Four or five cycles of
repressurization and pressure decay are
carried out in succession to ensure that
the compressor is running under constant
and reproducible conditions, as required
to obtain accuracy in the leak test
measurements.
TABLE 6. Pressure hold test data.
Date:
Time:
Average surface temperature, °C (°F)
Actual test pressure, kPa (lbf·in.–2 gage)
Final temperature corrected
test pressure, kPa (lbf·in.–2 gage)
Loss in test pressure, kPa (lbf·in.–2)
190
Leak Testing
Day One
_________________________
09:50
31.9 (89.5)
290 (42.0)
11:50
35.4 (95.8)
290 (42.0)
285 (41.4)
4.1 (0.6)
Day Four
________________________
15:30
32.2 (90.0)
290 (42.0)
17:30
32.6 (90.7)
290 (42.0)
289 (41.9)
0.7 (0.1)
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
The cyclic repressurization leak test
technique is based on the assumption that
no gas is supplied to the system under test
during the no load period. With a
reciprocating pump compressor unit,
when the load is removed it is not
uncommon for the compressor to
continue to deliver some gas during the
no load period. This can be avoided by
ensuring that the compressor is provided
with a delivery system that enables all gas
in the intercooler to be discharged into
the atmosphere during the no load
period.
Advantages and Limitations of
Cyclic Repressurization Leak
Testing
The cyclic repressurization leak testing
technique has the advantage of requiring
only very simple equipment. Its accuracy
in leakage measurements is less than the
accuracy of more direct leak testing
techniques. It is subject to random errors
caused, for instance, by malfunctioning
compressor valves. Therefore, these valves
should be checked for satisfactory
operation before starting each cyclic
pressurization leak test. The cyclic
pressurization test does not indicate the
volume of the system under test, nor does
it provide means for leak location.
When several compressors are
available, the compressor selected for the
leak tests should be one which if possible
gives charging times at least as long as the
leakage times. It is not advisable to
operate the compressor under part load
conditions, because its delivery capacity is
rarely determined with the same accuracy
for lower loads as for full load.
Compressors with dead space regulation
have a part load capacity that may differ
between (1) a compressor calibration test
and (2) a system leakage test, if the
quantity and temperature of the
compressor cooling water are different in
these two cases.
Localizing Leaks in Low
Pressure Gas Mains
Gas utilities use a modification of the
pressurizing mode of leakage
measurement to localize leaks in gas
mains. The main is tested, section by
section, with the leak locator inside the
pipe. The leak locator consists of the
following parts:
1. A flexible frame on which are spaced
two rubber gas bags jointed by a
rubber tube. These bags are pressurized
to seal off a short section of gas
pipeline for the leakage test.
2. A rubber dual tubing (two separate
tubular passages) of length sufficient
to reach from the rubber bags to a leak
test control panel. One of the tubular
passageways connects to the rubber
bags and is used for their
pressurization. The second tubular
passageway extends through the
adjacent rubber bag and opens into
the pipeline interior space between the
two rubber bag seals. This tube
transmits the contained natural gas
pressure to the control panel, where
any loss of pressure due to leakage of
gas from the test volume can be
monitored.
3. A control panel with an inclined water
gage connected to the test volume by
the rubber tube. This gage is used to
measure any variation in gas pressure
in the gas line section between the
rubber bag seals. A spring gage is used
to indicate the air pressure within the
sealing bags.
4. Connections are provided to a
pressure pump used to inflate the
rubber sealing bags and to a suction
pump that deflates the bags.
5. A steel rod is used to propel the bag
frame and tubing along the inside of
the gas main. The rod and bag frame
have sufficient flexibility to be passed
through a tape on the gas line and yet
have stiffness sufficient to avoid
buckling when the apparatus is
pushed along inside the gas main.
Procedure for Leak Testing of
Natural Gas Mains with Rubber
Sealing Bags
When testing for leakage in natural gas
mains, the sealing bags described
previously are inserted into a main
pipeline containing gas under moderate
distribution pressure. The rubber sealing
bags are spaced a set distance apart on the
frame. These bags seal off a portion of the
main line when they are inflated with a
pressure tire pump to 20 to 40 kPa (3 to
6 lbf·in.–2). The gas pressure in the test
volume between the two sealing bags is
indicated on the inclined water gage on
the test control panel as soon as the bags
are inflated to form pressure seals. When
the main gas distribution line is sealed off
completely by both bags, the pressure in
the test volume between the bags will
remain constant, as in a pressure hold
leak test. Loss of pressure would indicate
gas leakage through the wall of the
distribution pipeline, in the length
between the two sealing bags.
Pressure Change and Flow Rate Techniques for Determining Leakage Rates
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
191
PART 3. Pressure Change Tests for Measuring
Leakage in Evacuated Systems
Introduction to Pressure
Measurements in
Evacuated Systems
International System of
Units (SI Units) for
Vacuum Pressures
By popular usage, atmospheric pressure is
taken as the upper limit of vacuum. Any
pressure less than standard atmospheric
pressure (101 kPa) is some form of
vacuum. On Earth, vacuum pressure can
be anything between absolute zero
pressure and the barometer reading at the
particular location and time. Earlier, the
vacuum pressure was measured in inch of
mercy (in. Hg) or millimeter of mercury
(mm Hg) below atmospheric pressure. A
vacuum of 28 or 29 in. Hg was considered
to be a fairly good vacuum. Now, using SI
units, this same vacuum level would be
expressed as an absolute pressure of 3 to
6 kPa, which is 3 to 6 percent of normal
sea level atmospheric pressure, 101 kPa
(1 atm).
The SI unit for pressure is the pascal (Pa)
and is introduced here as the unit of
pressure in vacuums. Many processes
require medium levels of vacuum of the
order of 0.1 to 1 Pa. However, for many
applications such as high altitude
simulation chambers, pressures much
lower than 0.1 Pa are required. Units of
millipascal (mPa) or micropascal (µPa) are
used to describe pressures in this range of
hard vacuum, to avoid negative
exponents or powers of ten. The
previously used unit of torr (1 torr =
1 mm Hg) must be multiplied by 133 to
equal the pressure in pascal. The millitorr
is equal to pressure of 133 mPa. Because
the pressure of the standard atmosphere
at sea level is 1.01 × 105 Pa or 101 kPa, it
follows that perfect vacuum would have a
(negative) gage pressure of (–) 101 kPa
because the gage pressure in vacuum is
referred to the standard atmospheric
pressure at sea level.
Meaning of Absolute
Pressure and Gage
Pressure in Vacuum
Systems
As suggested earlier, the concept of a
vacuum is related to the pressure exerted
by the earth’s atmosphere. Atmospheric
pressure indicates the weight of a column
of atmospheric air of unit cross sectional
area measured at a particular altitude
above sea level. With increasing altitude,
the pressure decreases until, at some
indefinitely great height above the earth’s
surface (where only empty space exists),
the pressure approaches absolute zero. An
enclosure is said to be under vacuum if its
internal pressure is less than that of the
surrounding atmosphere. Because of
atmospheric pressure changes due to
meteorological factors and altitude, the
numerical value assigned to gage pressure
in vacuum is referred to atmospheric
pressure under standard conditions at sea
level (an absolute pressure of 101 kPa). As
vacuums were improved, it became
necessary to provide a scale of absolute
pressures (somewhat analogous to the
scale of absolute temperatures). The
concept of a perfect vacuum corresponds
to the hypothetical state of zero absolute
pressure.
192
Leak Testing
Conversions of Vacuum Pressures
from Prior Units to Pascal
The twentieth century has seen many
change in the units used to describe
pressure levels in vacuums. Early
investigators described their vacuum
pressure in terms of millimeter of
mercury, or torr, where the atmospheric
pressure at standard conditions was taken
as 760 torr. Hard vacuum pressures were
later described in terms of micrometer of
mercury (1 µm is one millionth of a meter
of mercury). Vacuum pressures are
variously expressed in pound per square
inch absolute pressure (lbf·in.–2 absolute),
inch of mercury, torr and the SI unit
pascal. For example, 1 µm Hg = 0.001 torr
= 10–6 m Hg = 133 mPa = 0.133 Pa. The
pressure of the standard atmosphere is
then equal to 760 torr. The (negative)
gage pressure for a perfect vacuum would
then be –760 torr in this system of units.
(An absolute pressure of 1 torr is equal to
133 Pa.) The preferred unit is pascal.
Conversion factors relate the various
units used to describe pressures in
evacuated systems, including the pascal,
atmosphere (atm or torr and the
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
micrometer of mercury). Figure 22 shows
a scale useful for approximate conversions
of vacuum pressures between SI units of
pascal and earlier units of millimeter of
mercury (equal to torr) for several typical
ranges of vacuum pressure. These
comparisons may also help personnel to
convert their data into SI units.
Limitations on Ultimate
Vacuum Pressure Caused
by Leakage and
Outgassing
During evacuation of a container,
molecules are constantly being removed
by the pumping process. Therefore, it
FIGURE 22. Histogram for conversion of vacuum absolute
pressures between prior unit of torr and SI unit of pascal
(1 std atm = 100 kPa = 760 torr).
133 kPa
1000
760
800
600
100 kPa = 1 atm
80
60
400
40
200
100
80
60
kPa
torr
20
10 kPa
8
6
might seem that eventually a pressure of
absolute zero would be obtained. This
would be true if the only molecules to be
removed were those in the gas space.
However, other gas sources do exist and
must be considered. The predominant gas
sources are leakage and outgassing.
Leakage is the direct transmission of gas
molecules, driven by the higher external
pressure, through holes or porosities in
the vacuum chamber wall, in welds or in
the various seals used in the system.
Outgassing refers to all forms of gas
coming from the materials in the vacuum
system. It includes gases that are adsorbed
on the surface, dissolved in the material
and occluded in gas pockets, as well as
those due to evaporation or
decomposition.
The continual addition of gas from
these sources represents the major
limitation on the ultimate pressure that
can be obtained in evacuated systems.
Mathematically, the ultimate pressure Pu is
given by the influx of gas Q divided by
the pumping speed S, so that Pu = Q /S.
Because the vacuum pump is itself a
source of outgassing, it can contribute a
limiting component Pp to the ultimate
vacuum pressure. Its effect is frequently
included in the prior equation for
ultimate pressure. In this case, Pu = Q /S +
Pp, where the term Q now refers to the
influx of gases from all sources except the
vacuum pump.
Even though the pump may be
operating at a particular limit pressure for
one type of gas, because of a leak or
outgassing, it can still pump other gases
to extremely low partial pressure. This is
true because, in molecular flow, all types
of gases flow independently of each other.
Typically, a gas analysis of an ultrahigh
vacuum system operating at a total
pressure of 10 nPa (~1 × 10–10 torr) will
show hydrogen and carbon monoxide as
the residual gases still coming from the
walls of the vacuum system in this
ultrahigh vacuum range. This occurs even
when the partial pressure of the original
nitrogen and oxygen are too low to be
measured.
40
Pumping Requirements for
High Vacuum Systems
4
25.4 torr = 1 in. Hg
20
2
10
8
1
The ideal gas laws apply to ideal gases
even at very low vacuum pressures. They
do not apply, however, to condensable
vapors such as water vapor or refrigerant
gases. The implications of the ideal gas
laws become evident when considering
the effect of reduced pressure on the
volume of a fixed quantity of ideal gas
held at constant temperature. A liter of
gas at standard atmospheric pressure
Pressure Change and Flow Rate Techniques for Determining Leakage Rates
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
193
would increase in volume as pressures are
lowered in the vacuum region (Table 7).
Tremendous multiplying factors come
into existence as the pressure drops in an
evacuated system. The pumping speed in
cubic meters per second does not increase
as pressure is lowered, so much smaller
masses of gas (fewer gas molecules) are
removed per unit of time, as system
pressure drops.
Residual Gas Molecule
Densities in High Vacuum
Systems
Because the mass reduction factor is so
great when evacuating a test system, it
might be assumed that after pumping to
reach low pressure, there is really nothing
in the container to affect any work that
may be inserted within it. However, one
must consider the number of molecules
that remain at various pressures. It may be
recalled that there is a physical
relationship stating that 22.4 L (0.89 ft3)
of any gas will contain 6.023 × 1023
molecules at 0 °C (32 °F) and 101.325 kPa.
The natural constant, 6.023 × 1023, is
known as Avogadro’s number. If the gas
pressure is now reduced to 0.1 Pa or one
millionth of its previous value, the 22.4 L
(0.89 ft3) volume of gas still within the
container contains 6 × 1017 molecules.
Even at 1 µPa, some 6 × 1012 molecules
will still remain in the 22.4 L (0.89 ft3)
volume. This provides a residue of
3 × 1011 molecules or almost one trillion
molecules per liter of volume (one billion
per cubic centimeter). To obtain low
ultimate vacuum pressures, one must
reduce the various sources of gas within
the system being pumped down. Leakage
can be eliminated only by first locating
each leak and then properly repairing it or
by placing an adequate temporary seal
over it. Maintaining cleanliness and
avoiding introduction of moisture into
the test system before the vacuum
pumpdown are vital. However, where
moisture has contaminated the interior
volume of a test system, vacuum pumping
can help to remove the moisture, if the
TABLE 7. Gas volume variation with pressure.
Gas Pressure
_____________
Pa
(atm)
105
103
100
10–3
10–6
10–9
194
(100)
(10–2)
(10–5)
(10–8)
(10–11)
(10–14)
Leak Testing
Volume of Gas
__________________
m3
10 –3
10 –1
10 2
10 5
10 8
10 11
(ft3)
(3.5
(3.5
(3.5
(3.5
(3.5
(3.5
×
×
×
×
×
×
10–2)
100)
103)
106)
109)
1012)
total amount of moisture is very small
(such as water adsorbed over a small
surface area).
Ensuring Cleanliness of
Welded Vessels to Be
Evacuated for Leak Testing
In preparation for leak testing by pressure
change or helium tests with a tracer probe
or hood, the interior of the system under
test is evacuated. A sensitive vacuum
pressure gage is then used to measure
pressure change or a helium mass
spectrometer is used to detect helium
tracer gas that reaches the vacuum pump
input. Joints for high vacuum vessels are
far more critical than joints in pressure
vessels that also operate under 100 kPa
(1 atm) of differential pressure.
Microporosity in the weld, entrapped
gases or solids and surface layers that
outgas become major problems with high
vacuum equipment or equipment that
will be evacuated for leak testing.
Extremely small defects or inclusions in
welded joints may not be detectable with
the usual nondestructive testing
techniques. The leak testing of the
evacuated system may be compromised
because of such small leak and gas
sources.
For valid leak detection and location by
the tracer probe technique or leakage
measurement by the hood technique,
cleanliness of the test object surfaces and
the leak testing system is essential. Tracer
gas can accumulate in surface dust and oil
or grease, including that within leak
passageways, possibly causing small leaks
to remain undetected when they are
exposed to the tracer probe gas only
briefly. Alternatively, evaporation of
condensed vapors and gases from such
contamination layers may cause a
sensitive leak detector system to indicate
leakage when the system is actually
leaktight. The larger the system under
test, the more important it is to ensure
cleanliness (including weld crevices and
surface discontinuities). The inert gas
tungsten arc welding (GTAW) process
produces clean welded joints with
minimum permeability to atmospheric or
tracer gases. The absence of welding flux
minimizes post weld cleaning operations
and problems of outgassing from slag
inclusions.
Description
atmosphere
high vacuum
very high vacuum
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
Effects of Weld Joint
Design on Leak Testing of
Evacuated Vessels
For pressure vessels to be evacuated
during leak testing (and vessels designed
for vacuum operation), the weld joint
design and preparation should avoid
trapped volumes or unwelded faying
surface areas that will be exposed to the
vacuum side of the joint. Both form
crevices that may hold foreign matter that
can outgas during evacuation or may
provide traps for tracer gases. Because
cleaning of such crevices is often
impossible, joint design and welding
procedures must eliminate such traps.
Welding should be performed from the
side of the joint that will be evacuated
whenever practical. The under bead often
contains unavoidable microporosity too
small to affect most strength and
toughness properties of the welded
structure. However, if exposed to the
vacuum, these voids could act as trapped
volumes. Leakage from this source can be
avoided by welding the cover (or seal)
pass from the side of the pressure
boundary that will be evacuated.
Figure 23a shows examples of preferred
joint designs for systems that will be
exposed to high vacuum. Figure 23b
shows undesirable joint designs which
provide dirt traps and create trapped
volumes (at the roots of butt welds made
from two sides of the plate, or fillet welds
with unwelded areas between abutting
plates).
Factors Influencing Speed
of Vacuum Pumping of
Large Volume Systems
The pumpdown time or time required for
evacuation of large vessels and systems
from atmospheric pressure is highly
dependent on the condition of the
vacuum system, the volume to be
evacuated and the pumping speed. Any
significant amount of water contained in
the system will have a powerful effect on
the time required for pumpdown because
water has a vapor pressure of 2.26 kPa
(17 torr) at 20 °C (68 °F). When water is
present within the system to be
evacuated, the pressure will not drop
below this value until the bulk of the
water has been pumped out. (Drying by
evacuation is often a useful way to
remove water trapped or condensed
within pressure vessels, piping and
components.) Consequently, water or
other vaporizing liquids should not be
introduced into test systems before leak
tests that require evacuation, if it can
possibly be avoided. Evacuation rates
attained by mechanical pumps drop
rapidly as the pressure is reduced by
FIGURE 23. Weld joint designs for welded vessels: (a) preferred designs have no crevices or volume traps open to evacuated
side of pressure boundary; (b) undesirable joints trap contamination and tracer gases, which may outgass during evacuation or
leak testing with sensitive mass spectrometer or other vacuum leak detectors.
(a)
(b)
T
T
T
T
T
Legend
= Vacuum side
= Continuous weld
T
= Locations of probable gas traps
= Intermittent weld
Pressure Change and Flow Rate Techniques for Determining Leakage Rates
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
195
pumping. Gas evolution by evaporation
of liquids at very low pressures increases
rapidly and prolongs the pumping period
required to attain desired vacuums.
Techniques for Estimating
Time Required for
Pumpdown to 10 kPa
(75 torr)
A technique for approximating the
mechanical pumpdown time for very
large systems as given by Guthrie in
Vacuum Technology7 uses the relation:
(64) T
=
2.3
V
S
where T is approximate pumpdown time
(2.3 time constants) to ten percent of
initial atmospheric pressure (to about
10 kPa or 75 torr); V is volume of test
system to be evacuated from atmospheric
pressure (100 kPa or 750 torr); S is
pumping speed of evacuation pumps,
volume unit per unit of time.
Consistent units must be used for each
term in the above equation, such as those
in Table 8.
Equation 64 indicates the pumpdown
time required to reduce pressure to one
tenth of an atmosphere or about 10 kPa.
To attain lower vacuum system pressures,
much more pumping time is required.
The term on the right side of Eq. 64 must
be multiplied by the factors in Table 9, for
various indicated final pressures within
the system being pumped down.
For example, to evacuate the system to
a pressure of only 1 Pa (7.5 mtorr), the
right side of Eq. 64 is multiplied by a
TABLE 8. Consistent units for pumpdown
calculation.
Time
Second
Second
Minute
Volume
Pumping Speed
Cubic meter
Liter
Cubic foot
Cubic meter/second
Liter/second
Cubic foot/minute
factor of 5, so that the pumpdown time is
estimated as:
T
=
2.3
×
5
V
S
=
11.5
V
S
Alternative Technique for
Estimating Pumpdown
Time to 10 kPa (75 torr)
An alternative approximation technique
for estimating pumpdown time of
practical industrial systems with prior
contamination is also presented by
Guthrie.7 This technique applies for many
average industrial systems that may have
various sources of gas, vapor and leaks
that will require larger pump sizes for any
given pumpdown time. For example, gas
may be trapped on interior surfaces by
mechanisms such as absorption (which
refers to binding of gas in the interior of
solid or liquid materials) or adsorption
(which refers to condensation of gas or
vapor on the surface of a solid). Despite
efforts to maintain or restore cleanliness
to the system, there will be variations
from system to system in the rates of
outgassing of these trapped gases and
vapors, which will change the required
pumpdown time to achieve specific
vacuum pressures.
This approximation technique makes
use of pumping down curves such as
would apply typically to clean systems of
known interior volume. For typical
industrial systems with contamination,
leaks or outgassing conditions, the time
indicated on the pumpdown curve for
clean systems would be multiplied by a
service factor that accounts for the effects
of nonideal systems. The service factors to
be used for average industrial systems are
listed in Table 10, in terms of the pressure
region to which the system will be
pumped down.
For example, suppose that a specific
system pumpdown curve shows a
pumpdown time of 200 min to pump
from 100 Pa (750 mtorr) to a final
pressure of 10 Pa (75 mtorr). The service
TABLE 10. Service factors for pressure
regions in pumpdown calculations.
TABLE 9. Calculation of approximate
pumpdown time.
Final Pressure
________________
196
Pa
Pa
(millitorr)
Multiplying Factor for
Equation 64
10
1
0.1
100
10
1
4
5
6
Leak Testing
Pressure Region
_________________________________
105
104
103
102
101
100
to
to
to
to
to
to
(torr)
104
103
102
101
100
0
(760 to 100)
(100 to 10)
(10 to 0.5)
(0.5 to 0.05)
(0.05 to 0.005)
(0.005 to 0.0002)
Service
Factor
1.0
1.25
1.5
2.0
3.0
4.0
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
factor corresponding to this final pressure
range is 2.0. The estimated pumpdown
time for this pressure range is then
obtained by multiplying this pumpdown
time of 200 min for an ideal system by a
factor of 2.0, to obtain an estimated
pumpdown time of 400 min in this
pressure range for the average industrial
system with contamination or leaks.
Comparison of Theoretical
and Actual Pumpdown
Curves for Welded Steel
Tank
The following is an example of a
theoretical and an actual pumpdown
curve for the annular space of a double
wall vacuum insulated liquified natural
gas tank. Figure 24 shows typical pump
curves relating pumping speed to pressure
for a combination of mechanical pumps
with booster pumps. For the test to be
reported here, the pump unit’s
performance curve is typical of several
shown in Fig. 24. Before the pumpdown
tests, the annular space welds were
deslagged. The metal surfaces were
examined with a near ultraviolet light
(used for fluorescent tracer inspection) to
detect any deposits of hydrocarbons
(which also fluoresce under ultraviolet
radiation). All deposits detected were then
removed with solvent cleaner to reduce
absorption of water vapor or other
condensable vapors within the interior of
the test volume. The entire interior
surfaces were then cleaned with a broom
to remove loose dust and dirt to eliminate
these particles as surfaces on which vapors
might condense and later outgas.
The technique for calculating
pumpdown time for very large systems is
used to predict the pumpdown time
periods ∆t between the initial pressure P1
and the final pressure P2 in accordance
with the approximation Eq. 65 and 66:8
V
S
(65)
∆t
=
K
(66)
∆t
=
2.3 K
=
K′
ln
V
S
P1
P2
P1
P2
log10
V
S
In Eq. 66, K’ = 2.3 K[log10 (P1/P2)]; ∆t is
the pumpdown time between the initial
pressure P1 and the final pressure P2.
Reasonable values for the K and for the K’
factors are given in Table 11.
K values cannot be added. However, for
calculating the pumpdown time ∆t for a
pressure range that spans two or more of
the pressure ranges listed above, the K’
value to be used is equal to the sum of the
K’ values given for the two or more ranges
covering the pressure difference for which
∆t is to be calculated. For example, if the
FIGURE 24. Curves relating pumping speed to pressure in vacuum chamber for various mechanical pumps with booster pump
units.
600
1 200
Instrument 1
500
Speed (L·s –1)
400
Instrument 3
800
Booster cut
in pressure
2.6 kPa (20 torr)
Instrument 4
Booster cut in pressure
2 kPa (15 torr)
300
600
200
400
100
200
Speed (ft3·min–1)
1 000
Instrument 2
0
0
10–2
10–1
100
(7.5 × 10–5) (7.5 × 10–4) (7.5 × 10–3)
101
102
103
104
105
(0.075)
(0.75)
(7.5)
(75)
(750)
Pressure, Pa (torr)
Pressure Change and Flow Rate Techniques for Determining Leakage Rates
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
197
pump speed is fairly constant from
atmospheric pressure (101 kPa) to a
pressure of 100 Pa (1 torr), an estimated ∆t
for the pressure range could be
determined in one calculation using a K’
value of 7.3 or (4.0 + 3.3).
The values of K and K’ for computing
pumpdown times as listed above apply
only for the case of clean mild steel tanks.
At pressures below 0.1 Pa (0.001 torr), the
pumpdown times are primarily
determined by outgassing conditions and
the relationships of Eqs. 65 and 66 are no
longer valid.
Vacuum Pumpdown
Technique for Leakage
Measurements
The evacuation pumpdown technique of
leak testing involves the determination
and evaluation of a pressure time response
curve for a vacuum test chamber within
which the test object is placed for testing.
Leakage measurement can be performed
in either of two ways.
1. Determining leakage rate at
equilibrium pressure attained during
pumpdown. The vacuum test chamber
is pumped down to equilibrium
pressure. Test object leakage and
outgassing from the test chamber are
measured and then subtracted from
the value of outgassing measured in a
leakfree system.
2. Deriving an allowable pressure time
curve for the pumpdown of a system
under test. Systems deviating from this
relationship are considered to be
leakers.
With either type of test system, it is
possible to set up an automated leak test
station involving a carousel system. The
carousel moves the test samples into
position, pumps them down and
measures the resultant pressures. The
biggest difficulty with this type of leak
test is the false reading produced by
outgassing of dirty samples.
TABLE 11. Values of pumpdown time estimation factors K
and K’.
Pressure Range
___________________________________
Pa
101 000 to 2600
2600 to 133
133 to 13.3
13.3 to 6.6
6.6 to 1.3
1.3 to 0.13
198
Leak Testing
(torr)
K
K’
(760 to 20)
(20 to 1.0)
(1 to 0.1)
(0.1 to 0.05)
(0.05 to 0.01)
(0.01 to 0.001)
1.1
1.1
1.5
4.0
4.0
4.0
4.0
3.3
3.45
2.77
6.44
9.21
Equations Used in Analysis of
Vacuum Pumpdown Leak Tests
The fundamental response curve for a
vacuum system during pumpdown is
described by Eq. 67:
(67)
dP
dt
Q
V
=
−
S
P
V
where P is pressure in system being
evacuated; t is time elapsed from start of
pumping; S is effective pumping speed of
vacuum pump; V is volume of system
being evacuated; Q is total in-leakage rate
plus outgassing load of test system; dP/dt
is time rate of change of pressure. The gas
load may be due to leakage, evolution of
gas from the walls of the evacuated
system or both.
In the lower vacuum pressure range
where outgassing has significant effect,
integration of Eq. 67 leads to the
pumpdown response characteristic of
Eq. 68:
(68) t 2 − t 1
=
−
V
S
1 − P2
ln
1 − P1
S
Q
S
Q
Equation 68 describes an exponential
decay curve with a time constant equal to
S/V, which becomes asymptotic to an
equilibrium pressure defined by Eq. 69:
(69)
P
=
Q
S
This ultimate pressure is approached in
approximately five time constants, when
t2 – t1 = 5(S/V).
Procedure for Pressure Rise
(Vacuum Retention) Test for
Leakage Rate
The pressure rise test (also called a
vacuum retention test) is a pressure
change leakage measurement technique
performed on a system evacuated below
atmospheric pressure. It can be performed
on systems at any vacuum level but is
most effective on systems evacuated to an
absolute pressure (vacuum) in the range
from 10 to 0.001 Pa (100 to 0.01 mtorr).
This leakage rate test is performed by
isolating the system under test after it has
been evacuated to the required (or
specified) absolute pressure (vacuum).
Then the pressure and, when exposed to
ambient weather conditions, the surface
temperature of the system are observed
for a specific time to determine the rate of
pressure rise per unit of time for the
system. Figure 25 shows schematically the
test arrangement and the connections
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
FIGURE 25. Arrangement of equipment for pressure rise leakage rate testing of an evacuated
system. Also known as a vacuum retention test, this test measures overall leakage rates and
requires use of a vacuum pumping system and a vacuum gage. For systems exposed to
ambient weather conditions, surface temperature detectors are used to approximate internal
air temperatures in the system. Ambient temperature must be measured in shade, not in
direct sunlight.
Surface thermometer
Boundary of test system
Evacuated
system
Surface
thermometer
Surface thermometer
Closed
during test
Open during test
Vacuum pump
system
Gage tube
Vacuum gage
Optional valve
between the test volume, the vacuum
pump system and the instrumentation.
Effects of Condensable
Vapors on Vacuum
Retention Leakage Test
As noted earlier in this chapter, the
behavior of vapors in an evacuated system
deviates significantly from the General
ideal gas law:
(70)
PV
=
n RT
A vapor is the gaseous form of any
substance that usually exists in the form
of a liquid or solid, such as water vapor. A
pure liquid in equilibrium with its own
vapor will have two phases (liquid and
vapor) that coexist at a specific partial
pressure known as the vapor pressure.
Because condensation or evaporation
occurs with changes in temperature, vapor
molecules enter or leave the gaseous
phase with any change in temperature.
This changes the number of molecules of
a particular vapor and the partial pressure
which that vapor exerts in a particular gas
volume. These vapor effects, called
outgassing in a vacuum system, are not
included in the effects described by the
General ideal gas law of Eq. 70. For this
reason, in an evacuated system, it is not
mathematically realistic to make accurate
temperature corrections to the final
pressure for pressure data taken at
different temperatures.
Therefore, to establish a fairly accurate
leakage rate by this pressure change
technique for an evacuated system
exposed to ambient weather conditions, it
is necessary to compare pressure data at
periods when the temperature is the same
or nearly the same and the temperature
trends are in the same direction. For a
system enclosed in a temperature
controlled building, such as a vacuum
chamber evacuated to lower absolute
pressure ranges, temperature
measurements are usually not necessary. A
pressure rise test of such an enclosed
system can be used to determine both the
leakage rate and the outgassing rate for
that system.
Advantages of Pressure
Rise (Vacuum Retention)
Leakage Test Technique
The pressure rise leakage rate test is
relatively simple in principle and fairly
Pressure Change and Flow Rate Techniques for Determining Leakage Rates
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
199
easy to perform on smaller test systems.
The test is capable of attaining increased
leakage sensitivity as the system size or
volume decreases. That is, the total
leakage rate that can be measured as a
pressure rise per unit time becomes
smaller as the system under test gets
smaller in volume. This test technique can
serve as a final test or as a preliminary test
preceding other leak test techniques,
depending on the size and configuration
of the system to be leak tested. This
quantitative leakage rate test can be used
to determine the total leakage rate (in the
form of a pressure rise per unit of time)
through the test boundary of any system
capable of being evacuated.
Limitations of Pressure
Rise (Vacuum Retention)
Leakage Test Technique
The sensitivity of the pressure rise leakage
rate test diminishes as the size or volume
of the system to be tested increases. Larger
rates of leakage must exist if they are to
be detected in large volume systems by
this test technique. In addition, the
location of unacceptable leakage cannot
be determined by this test alone. If the
actual total leakage rate exceeds the
allowable value, another leak test
technique must be used to locate any
unacceptable leaks or the numerous small
leaks that might contribute to an
unacceptable high overall rate of leakage.
Thus, performance of a pressure rise test
on the evacuated annular space of a
double wall vessel, with a resultant total
leakage rate indication in excess of that
allowable, will not reveal whether the
unacceptable leakage is in the inner
vessel, in the outer vessel or in a
combination of both.
Because of the effect of vapors that do
not obey the general gas laws for ideal
gases, it becomes difficult to determine an
accurate true gas pressure rise per unit of
time for very large volume systems
exposed to wide temperature variations
during the leakage test period. Lowering
the absolute pressure within the
evacuated vessel in an effort to increase
the leakage rate test sensitivity may be
unfeasible because of the vacuum
pumping system limitations. Alternatively,
the rate at which gas can be pumped out
may be limited by the size of the hole
(penetration) through which it must be
removed. Trying to increase the test
sensitivity by increasing the duration of
the test, in an effort to achieve the ability
to read a smaller pressure rise per unit
time more reliably, may prove unrealistic
as costs increase and schedule completion
is made more difficult.
200
Leak Testing
Factors Affecting Leakage
Sensitivity of Pressure Rise
Test Technique
The leakage rate sensitivity of the pressure
rise (or vacuum retention) leakage rate
test is influenced by five major factors:
1. absolute pressure attained in the
evacuated system, when the test is
performed (this, in turn, affects the
resolution of the smallest measurable
pressure change);
2. internal volume of the system to be
tested;
3. time duration of the leakage rate test;
4. ambient temperature and weather
conditions; and
5. internal surface areas and cleanliness
of the test system.
Each of these factors is discussed next, in
greater detail.
Effect of Absolute Pressure in
Evacuated System Being Tested
When vacuum retention leakage rate tests
are performed within the absolute
pressure range of 10 to 0.001 Pa (100 to
0.01 mtorr) on large systems, the lower
the pressure, the greater the test
sensitivity becomes. The limitation on the
high pressure end of this range results
from inability to measure very small
pressure changes resulting from leakage
from large volumes.
For example, it might be necessary to
detect changes of a fraction of a pascal at
2.5 kPa (a few micrometers at 20 torr).
The limitation on the low pressure side is
the increase in the portion of the pressure
change attributable to outgassing. At these
very low absolute pressures, the pressure
rise due to actual leakage is small in
relation to the pressure rise due to
outgassing. This makes it difficult to
determine the true rate of pressure rise
caused by real leakage.
Effect of Volume of Tested System
The test sensitivity and, in turn, the rate
of pressure rise both vary inversely with
the size or volume of the evacuated system
being tested. For example, a leakage rate of
5 × 10–3 Pa·m3·s–1 (5 × 10–2 std cm3·s–1) in
a 570 m3 (2 × 105 ft3) system would cause
a rate of pressure rise of only 0.8 Pa
(5.8 mtorr) per day. This same rate of
leakage in a 0.3 m3 (10 ft3) system would
cause a rate of pressure rise of 1.5 kPa
(11.6 torr) per day.
Effect of Duration of Leakage Test
The sensitivity of the leakage rate test
increases directly with the elapsed time
during the test. As the time duration of
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
the test increases, the test sensitivity
increases. The three factors of absolute
pressure P, system volume V and time
duration t of the pressure rise test are
related by Eq. 71 and 72:
(71) Q
=
(P2
− P1)
V
t
where Q is leakage rate (Pa·m3·s–1); P1 is
initial absolute pressure (torr); P2 is final
pressure (pascal); V is volume of
evacuated system under test (cubic meter);
t is time duration of test (second).
(72) Q
=
(P2
− P1 )
V
96.6 t
where Q is total leakage rate (std cm3·s–1);
P1 is initial absolute pressure (torr); P2 is
final absolute pressure (torr); V is volume
of evacuated system (cubic foot); t is time
duration of test (hour). For other systems
of units, the conversion factor of 96.6 will
change.
Effects of Weather and Ambient
Temperature Conditions
In pressure rise (vacuum retention) tests
of evacuated systems, the greater the
exposure of the system to direct sunlight
and the greater the variations in ambient
temperature, the more difficult it becomes
to determine an accurate pressure rise.
Temperature variations lead to
uncontrollable effects on the rate of
outgassing or condensation of vapors
within the system, which also influence
the pressure variations in the system.
Effects of Internal Surface Area and
Cleanliness of Test System
With evacuated systems under pressure
rise leak testing conditions, the smaller
the internal surface area and the cleaner
that surface is, the less the outgassing in
the systems. This reduces the effect on
pressure change from outgassing due to
temperature variations.
Estimating Leakage Test
Sensitivity Attainable in
Pressure Rise Tests
To determine the leakage rate sensitivity
attainable with a pressure rise (vacuum
retention) test, it is necessary to know in
advance the volume (estimated or
calculated) of the system and the absolute
pressure at which the test must be
performed. If the allowable pressure rise
per unit of time is known or specified and
it is realistic for the absolute pressure
(vacuum) level at which the test is to be
performed, the test sensitivity or
detectable total leakage rate can be
computed by Eq. 71. If instead the test
sensitivity or total leakage rate Q is
specified or known because of system
performance requirements, the allowable
pressure rise per unit of time for that total
leakage rate can be determined by using
the transposed form of Eq. 71 shown
below as Eq. 73:
(73)
P2 − P1
t
=
Q
V
=
96.6
or in torr·h–1:
(74)
P2 − P1
t
Q
V
Units for variables in Eqs. 73 and 74 are
given below Eqs. 71 and 72, respectively.
If the rate of pressure rise computed by
Eqs. 73 and 74 is measurable with
available test equipment at the specified
test pressure, the required or specified
leakage rate test sensitivity can be
achieved. If it is not measurable, then an
attainable test sensitivity must be
established by Eqs. 71 and 72.
Example Computation to
Determine Pressure Rise Test
Feasibility
As an example of the application of
Eq. 74, suppose that the performance
specification for a system requires that the
completed system contain no leakage in
excess of 2 × 10–3 Pa·m3·s–1 (2 × 10–2 std
cm3·s–1). This 300 m3 (105 ft3) system can
be evacuated to an absolute pressure of
1 Pa (or about 10 mtorr) with the
permanent vacuum pump system. Would
a pressure rise test be a realistic test
technique for quantitatively verifying that
this system meets the specification
requirements? Because Q and V are
known, Eq. 73 can be solved as follows in
SI units: (P2 – P1)/t = Q/V =
(2 × 10–3)/300 = 7 × 10–6 Pa·s–1 or 0.6 Pa
per day. For mixed units, Eq. 74 indicates
that:
P2 − P1
t
=
96.6
=
96.6
=
Q
V
2 × 10 −2
10 4
1.93 × 10 −2 torr ⋅ h −1
These requirements can be met by the
pressure rise (vacuum retention) leakage
rate test technique. The time required for
the test depends on the surrounding
temperature conditions. If the system is in
a building in a controlled temperature
environment, a test duration of only a
few hours should be adequate. If the
system is exposed to the weather, then a
Pressure Change and Flow Rate Techniques for Determining Leakage Rates
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
201
The specification for the pressure rise
(vacuum retention) leakage rate test
required that the test be conducted over a
period of 72 h. The allowable pressure rise
was 3.3 Pa (25 mtorr) in 72 h. For this
time span, this was a reasonable leakage
allowance. For the 650 m3 (2.3 × 104 ft3)
annulus volume and allowable pressure
rise rate of 3.3 Pa (25 mtorr), the total
leakage rate allowable was computed in SI
units as:
comparable temperature cycle must be
experienced. If the weather is cloudy and
the temperature is stable, a few hours may
be adequate. Normally, for an exposed
system, a temperature cycle of 12, 24 or
36 h is necessary to achieve the necessary
reliable comparison data.
Example of Pressure Rise Leakage
Rate Test of Liquid Hydrogen
Vessel
The following example illustrates test
conditions and test results for the leakage
rate of the annular inner space between
concentric inner and outer spheres of a
double wall vacuum insulated liquid
hydrogen vessel. The outer sphere has a
15.81 m (51 ft, 10.5 in.) inside diameter
and the inner sphere has an inside
diameter of 13.9 m (45 ft, 7 in.) and a
wall thickness of about 19 mm (0.75 in.).
The volume of this annular space was
calculated to be about 650 m3
(2.3 × 104 ft3). Critical areas of the inner
sphere were tested by the more sensitive
helium tracer probe or hood leak testing
techniques.
Q
=
3.3 × 650
72 × 3600
=
8.3 × 10 −3 Pa ⋅ m 3 ⋅ s −1
=
8.3 × 10 −2 std cm 3 ⋅ s −1
The results of the pressure rise test
performed on the annular space of this
double wall liquid hydrogen sphere are
shown in the pressure rise test data of
Table 12 and are plotted in the graphs of
pressure and temperature as a function of
time during testing in Fig. 26. Pressure
levels may be compared at any of the
nearly equivalent temperature points
during the night time periods marked
Absolute pressure, Pa (mtorr)
Temperature, °C (°F)
FIGURE 26. Graphs showing variations in temperature and absolute pressure of liquid hydrogen sphere annular space during
72 h pressure rise leakage rate test. Arrows with asterisks indicate time periods when temperatures and trends in change of
temperature were comparable. Pressure rise test liquid hydrogen sphere with 13.9 m (45 ft, 7 in.) inside diameter inner tank
and 15.8 m (51 ft, 10.5 in.) inside diameter outer tank.
43
(110)
38
(100)
32
(90)
27
(80)
21
(70)
16
(60)
6.7
(50)
5.3
(40)
4.0
(30)
2.7
(20)
1.3
(10)
Average shell
temperature
Ambient
temperature
Absolute
Pressure
0
600
1000
1400 1800
2200
200
600
1000 1400 1800 2200
200
600
44
48
1000 1400 1800 2200
200
600
68
72
Real time (h)
0
4
8
12
16
20
24
28
32
36
40
52
56
60
64
Elapsed time (h)
202
Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
with brackets and footnotes on the
elapsed time column of Table 12. These
time points are also marked by arrows and
asterisks on the graphs of Fig. 26. The
results for this 52 h time span indicate
that the pressure rise on this system could
have been a maximum of 250 to 400 mPa
(2 to 3 mtorr) in 72 h. This was an
acceptable leakage test rate because it was
much less than the allowable rate of 3 kPa
(25 torr) in 72 h. The total leakage rate is
equivalent to 6.6 × 10–4 to 9.9 × 10–4
Pa·m3·s–1 (6.6 × 10–3 to 9.9 × 10–3 std
cm3·s–1). Because this loss of 250 to
400 mPa (2 to 3 mtorr) is less than the
error in reading of the Pirani or
thermocouple vacuum pressure gage used
for the test, the pressure rise was probably
much less. A longer test period could have
proved this but would have served no
useful purpose.
Example of Pressure Rise Leakage
Rate Test of Laboratory Vacuum
Chamber
The following example illustrates test
conditions and test results for a pressure
rise test of a stainless steel solvent cleaned
vacuum chamber. The purpose of the test
was to determine the leakage rate and the
outgassing rate of the chamber. The
chamber had an inside diameter of
590 mm (23.25 in.) and a length of 1.6 m
(63 in.). Its volume was calculated to be
about 0.487 m3 or 487 L (17.2 ft3). Its
inside surface area was calculated to be
about 4.63 m2 (49.8 ft2). The results of the
pressure rise test with the chamber
vacuum conditioned for approximately
189 h are given in Table 13 and shown
graphically in Fig. 27. Figure 27b is an
enlargement of the upper linear portion
of the graph of Fig. 27a, from whose slope
the final leakage rate was determined.
TABLE 12. Test data for pressure rise test of liquid hydrogen sphere.
Real
Time
Elapsed
Time (h)
0600
0800
1000
1200
1400
1600
1800
2000
2200
2400
0200
0400
0600
0800
1000
1200
1400
1600
1800
2000
2200
2400
0200
0400
0600
0800
1000
1200
1400
1600
1800
2000
2200
2400
0200
0400
0600
0
2
4
6
8
10
12
14
16b
18b
20b
22
24
26
28
30
32
34
36
38
40b
42b
44b
46
48
50
52
54
56
58
60
62
64
66
68b
70b
72b
Shell Temperature, °F a
________________________________
Annulus
Press
____________
No. 1
Pa (millitorr)
59
65
72
78
81
83
79
70
65
63
61
58
56
63
76
80
82
84
78
67
63
61
59
57
56
65
82
85
90
93
87
80
73
70
65
62
59
No. 2
No. 3
Ambient
Temperature
Average
°F
56
77
100
112
100
97
87
69
65
62
59
57
53
81
108
120
113
106
86
70
65
62
60
58
55
87
120
129
124
120
110
93
78
72
68
63
59
57
62
69
81
100
97
82
69
63
61
59
56
54
57
69
85
96
98
80
68
63
61
58
56
54
70
100
103
106
104
97
86
75
66
62
61
59
57.3
68.0
80.3
90.3
93.7
93.0
82.7
69.3
64.3
62.0
59.7
57.0
54.3
67.0
84.3
95.0
97.0
96.0
81.3
68.3
63.7
61.3
59.0
57.0
55.0
74.0
100.7
105.7
106.7
105.7
98.0
86.3
75.3
69.3
65.0
62.0
59.0
60
66
72
77
80
80
78
69
65
63
61
59
57
65
77
79
80
80
78
68
64
62
60
58
57
71
80
85
87
88
85
80
73
68
65
63
60
2.0
2.7
3.9
5.1
6.6
6.9
6.3
3.7
2.7
2.4
2.1
1.9
1.5
2.4
3.5
4.5
5.7
6.0
5.3
3.5
2.5
2.3
2.1
1.7
1.3
4.0
8.0
10.1
10.7
10.7
10.0
8.1
6.0
3.9
2.9
2.5
2.3
(15)
(20)
(29)
(38)
(49)
(52)
(47)
(28)
(20)
(18)
(16)
(14)
(11)
(18)
(26)
(34)
(43)
(45)
(40)
(26)
(19)
(17)
(16)
(13)
(10)
(30)
(60)
(76)
(80)
(80)
(75)
(61)
(45)
(29)
(22)
(19)
(17)
Information and
Comments
Begin hold test
Windy, clear and sunny
Clear and calm
Clear, calm and sunny
Clear and calm
Clear, calm and sunny
End of 72 h hold test
a. (°F – 32)/1.8 = °C.
b. Data comparison points.
Pressure Change and Flow Rate Techniques for Determining Leakage Rates
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
203
The total rate of pressure rise due to
both outgassing and leakage during the
entire 41 h, 44 min (≅ 2500 min =
150 000 s) test period was computed as:
Total Q
(1.64
×
=
7.0 ×
10 −6
=
7.0 × 10 −5 std cm 3 ⋅ s −1
=
) (1.72
×
Pa ⋅ m 3
⋅ s −1
10 −2
TABLE 13. Pressure rise leakage rate test of type 487 L
stainless steel vacuum chamber. See Fig. 27.
101
96.6 × 41.73
)
Real Time
(h : min)
16:36
16:38
16:39
16:40
16:41
16:44
16:47
16:53
16:57
08:50
10:50
14:00
16:50
08:30
10:20
Based on the straight line portion of
the last 23.5 h of the test as shown in the
graph of Fig. 27b, the leakage rate Q for
the chamber was computed as:
FIGURE 27. Pressure rise leakage rate test of a type 487-L
stainless steel vacuum chamber: (a) pressure rise as a
function of time; (b) enlargement of upper portion of curve,
showing rate of pressure rise due to leakage, following
outgassing of steel vacuum chamber. See Table 13.
Pressure, Pa (torr)
(a)
1.3 × 100
(10–2)
1.3 × 10–1
(10–3)
1.3 × 10–2
(10–4)
1.3 ×
(10–5)
Q
=
1.3 × 10–4
0
0.033
0.05
0.067
0.083
0.133
0.183
0.283
0.35
16.23
18.23
21.40
24.23
39.90
41.73
(8.4
× 10 −3
2.9
7.2
9.5
1.2
1.3
1.9
2.7
4.1
4.9
1.0
1.1
1.2
1.3
2.1
2.2
×
×
×
×
×
×
×
×
×
× 101
96.6 × 23.5
6.4 × 10 −6 Pa ⋅ m 3 ⋅ s −1
=
6.4 × 10 −5 std cm 3 ⋅ s −1
10–4
10–4
10–4
10–3
10–3
10–3
10–3
10–3
10–3
(torr)
(2.2 × 10–6)
(5.4 × 10–6)
(7.1 × 10–6)
(8.7 × 10–6)
(1.0 × 10–5)
(1.4 × 10–5)
(2.0 × 10–5)
(3.1 × 10–5)
(3.7 × 10–5)
(7.2 × 10–3)
(8.0 × 10–3)
(8.7 × 10–3)
(9.7 × 10–3)
(1.58 × 10–2)
(1.64 × 10–2)
)
(10–6)
10
20
30
40
50
=
6.0 × 10 −7 Pa ⋅ m 3 ⋅ s −1
=
6.0 × 10 −6 std cm 3 ⋅ s −1
=
4.6 × 10 −6 torr - L ⋅ s −1
This results in outgassing computed in
torr-L·s–1·cm–3 as:
Elapsed time (h)
Outgassing
(b)
Absolute pressure, Pa (torr)
Pa
Subtracting the leakage rate from the total
rate results in the outgassing rate
computed as:
0
=
=
1.5 × 10 –2
) (1.72
Chamber Pressure
__________________________
=
Q
10–3
Elapsed
Time (h)
(2)
4.6 × 10 −6
4.63 × 10 4
1.0 × 10 −10
per square meter of surface area. This
agrees very closely with published
outgassing data for degreased stainless
steel with 200 h vacuum conditioning.
Leakage rate
1 × 10 –2
(1.3)
5 × 10 –3
(0.7)
After 10:50
0
4
8
12
16
20
24
Elapsed time (h)
204
Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
PART 4. Flow Rate Tests for Measuring Leakage
Rates in Systems near Atmospheric Pressure
Principles of Leakage
Testing by Measurement
of Flow Rates
The flow measurement procedure for
leakage testing consists of determining the
extent of leakage by measuring the rate of
flow of gas moving into or out of the
system or component under test. Flow
rates can be measured with a flow meter
or by means of pumping at known
volumetric pumping rates to maintain a
fixed system pressure or to compare rates
of change of pressure. The flow
measurement leakage test procedure can
be roughly separated into two broad
classes of technique: (1) observation and
measurement of gas flow rates or volume
of gas displaced and (2) analysis of effects
of pumping gas during pressurization or
evacuation of systems, on pressure or rates
of change of pressure. When leak testing
by the flow observation technique, the
amount of leakage is measured. The
system under test is pressurized or
evacuated and placed within a sealed
enclosure. The enclosure volume is
connected through a flow meter to a
regulated pressure source. The gas transfer
by leakage between the system under test
and its enclosure causes a pressure
difference between the enclosure volume
and the regulated pressure source. The gas
transfer between the sealed enclosure and
the reference pressure source is measured
by flow meters, by movement of a liquid
(slug) indicator in a capillary tube in
which the leaking gas is accumulated or
by other techniques. In some cases, the
reference pressure may be atmospheric
pressure. Figure 28 shows a leakage testing
system using a fluid slug indicator of the
amount of gas leakage.
Pumping Technique for
Measuring Leakage Rate
from Evacuated Test
Systems
In the pumping technique of leakage
testing of evacuated systems, the system
under test is evacuated by a vacuum
pump. The rate of system pressure
decrease during pumpdown is then
compared with the rate of pressure
decrease during pumpdown of a leaktight
system.
In an alternative leak testing
procedure, the sealed enclosure can be
evacuated and allowed to reach pressure
equilibrium with its vacuum pumps. The
rate at which gas is being pumped to
maintain this equilibrium is then
measured to determine the rate of leakage
from the test volume into the enclosure.
Pumping Technique for
Measuring Leakage Rate
from Pressurized Systems
In an alternative pumping technique for
measuring leakage rates, the test volume
can be pressurized and the compressor is
then operated only sufficiently to keep
the test system pressure constant. The
leakage rate can then be calculated from
the volumetric pumping speed (m3·s–1)
and the length of time the compressor
must operate to regain a predetermined
system pressure.
Sensitivity of Flow
Measurement Leak Testing
Techniques
The sensitivity of leakage rate testing by
flow measurements is relatively low,
compared to the sensitivity of many other
leak testing techniques described in this
volume. In most cases, the leakage
sensitivity depends on that of the
FIGURE 28. Arrangement for leakage rate testing of system
enclosed in a sealed test enclosure connected to a capillary
tube flow meter with an opaque visible liquid indicator slug.
Leakage from pressurized system into enclosure would cause
indicator slug to move to the right by a displacement
proportional to the volume of gas leakage.
Liquid
indicator
slug
System
under
test
Enclosure
Connection to
reference volume
or pressure source
(or atmosphere)
Capillary tube
Pressure Change and Flow Rate Techniques for Determining Leakage Rates
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
205
instrument used to measure the flow rate
and is relatively independent of the test
system volume. In a flow observation
technique, leakage rates between 10–3 and
10–5 Pa·m3·s–1 (10–2 and 10–4 std cm3·s–1)
can be detected, depending on the flow
instrument used. If a sealed system is
being evacuated, flow rates of the order of
0.1 Pa·m3·s–1 (1 std cm3·s–1) may be
observed. (Note that 1 Pa·m3·s–1 is
equivalent to 10 std cm3·s–1.)
The leakage sensitivity attainable with
the pumping pressure analysis technique
depends on the size (pumping speed ) of
the pumps. With evacuated test objects or
test systems, leakage sensitivity depends
critically on the outgassing within the
system being measured.
Advantages and
Limitations of Leak Testing
by Flow Measurements
Flow measurement leak testing procedures
are applicable to a large variety of test
systems. The procedures are useful only
for measurement of leakage. They are not
appropriate for locating leaks. They are
used to measure total leakage rates in
small sealed parts. They can be used to
measure total leakage rates in large sealed
systems and in systems that can be
pressurized or evacuated. The major
advantages of leak testing by means of
flow measurements are as follows.
1. No special tracer gas is necessary. the
flow measurement leak testing
procedure is applicable to whatever
fluid is present within the system to
be tested. The test system need not be
placed in any special environment for
leak testing. Instead, systems may be
tested in their normal operating
modes.
2. The cost of the equipment for flow
measurement leak testing is low.
3. The sensitivity of overall leakage
measurement is independent of
system volume.
4. The leakage rate can be measured
without extensive calibration.
However, the accuracy of leakage
measurement is not very high, as
compared with that for many other
techniques.
5. When calibration is required, it can be
readily attained with standard flow or
volume measurement equipment.
There are two major disadvantages of
flow measurement leak testing.
1. The test sensitivity is low.
2. Flow measurement procedures have
not gained wide recognition.
Flow measurement uses various types of
equipment with little similarity and
206
Leak Testing
different techniques are used to solve
individual leak testing problems.
Sealed Volume Technique
of Leak Testing by Flow
Measurements
Figure 28 shows the arrangement of leak
testing equipment using the most
common technique of flow measurement
by observation of the movement of fluid
in a glass capillary tube. The system under
test is enclosed and sealed within the test
enclosure. The system being tested can be
either evacuated or pressurized. It can
either be sealed or connected to a source
of pressure or of vacuum.
Care must be taken to ensure that the
leakage being measured is not occurring
in the connection to the source of
pressure or vacuum. The capillary
containing the indicating fluid is attached
to the test enclosure. This type of testing
can be performed with the capillary fluid
indicator connected between the test
enclosure and a standard testing volume
on the other end of the capillary. In this
way, the leak test can be compensated for
temperature variations, if both test
enclosure and the comparison volume are
subject to the same temperature
conditions. Alternatively, the capillary can
be connected between the test enclosure
and the atmosphere. For accurate leakage
measurements and rapid response, the
enclosure containing the system under
test should have a net volume as small as
practical.
One advantage in the construction of
the sealed volume type of leak testing
equipment shown in Fig. 28 is that there
are no critical, leaktight connections
within the enclosure. This is because the
system is operating at atmospheric
pressure. Therefore, although it is possible
that a leak could exist between the
enclosure and the atmosphere, leakage
does not occur through this leak because
no pressure differential is applied across it.
Any differences in pressure are
compensated for by the pressure
transmission through the liquid slug
within the interconnecting capillary tube.
Measuring Leakage Rates with Glass
Capillary (Pipette) Tubes
Glass capillary tubes containing a slug of
indicating fluid provide a means for direct
quantitative measurement of leakage rates
if a record is made of the time required
for the small liquid plug to move a given
distance. Because the cross sectional area
of the capillary bore is known, the
volume swept out by the liquid plug
during the measured time interval can be
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
computed. A 1.5 mm (0.06 in.) diameter
glass capillary tube is used to measure
leakage rates in the range from 10–3 to
10–1 Pa·m3·s–1 (10–2 to 100 std cm3·s–1). A
0.5 mm (0.02 in.) glass capillary tube can
be used to measure smaller leakage rates
from 10–5 to 10–3 Pa·m3·s–1 (10–4 to
10–2 std cm3·s–1). These capillary tubes are
marked with scales given in convenient
units for computing leakage rates. A
stopwatch is commonly used for timing
the movements of the liquid plug within
the capillary tube. Pipettes used for liquid
measurements provide convenient
calibrated capillary tubes.
The upper limit on leakage rates
measurable with capillary tubes is reached
when the liquid plug moves so fast that
timing is difficult. The lower limit on
leakage rate measurement is determined
by the accuracy desired and is influenced
by errors introduced by the resistive and
inertial forces affecting the movement of
the liquid plug within the capillary tube.
Changes in atmospheric pressure
(barometric readings) may move the
liquid slug in capillary systems with one
end open to the earth’s atmosphere. As
the speed of movement of the liquid plug
decreases, these errors are increased. This
causes the leakage measurements to
become more inaccurate with slow
movements of the liquid plug.
Errors due to starting inertia are
decreased with liquids of lower density.
Errors can be reduced, for example, by
using a water plug about 1 mm (0.04 in.)
long and timing the movement of the
water plug only after it reaches a constant
velocity. If a water plug is used, the error
due to the resistive forces of surface
tension can be minimized by coating the
inside (bore) surface of the clean capillary
tubing with an organosilicon compound.
This coating acts to prevent the water
from wetting the glass.
Mercury is almost impossible to use for
the liquid slug in a glass capillary.
Mercury has a very high surface tension
and it is almost impossible to force it into
a very small diameter capillary tube bore.
However, there should be negligible gas
transfer through a mercury plug.
An ideal fluid for use as the indicator
plug in a glass capillary should have the
following characteristics.
1. It should be a fluid in which the
leaking gas is not soluble, so that no
gas transfer by diffusion can occur
through it,.
2. The fluid should not wet the walls of
the tube, so that the surface tension
forces on either end of the plug are
balanced.
3. The fluid should be opaque for easy
visibility and measurement of its
position.
4. The fluid should have a low surface
tension so that it can be placed easily
within the bore of the capillary tube.
Alternative Flow
Measurement Instruments
Used in Sealed Volume
Leakage Tests
The basic principles of sealed volume leak
testing can be used in numerous ways. For
large leaks, flow measuring devices such
as a wet type gas meter or a rotameter
may be used. These instruments produce
accurate leakage rate measurements but
are useful only on very large leaks. For
measurements over a wide range of
leakage rates, the instrument shown in
Fig. 29 can form a U-tube capable of
withstanding extremely high pressures.
Tubes B and C have different diameters so
that the proper tube can be selected for
measuring various leakage rates.
When all the valves in Fig. 29 are open
and the test components are pressurized,
the liquid columns all reach the same
height. By closing the shutoff valve in the
main line between columns A and B,
leakage is indicated by upward movement
of fluid in columns B or C. The meter in
Fig. 29 was designed primarily for
determining leakage rates in hydraulic
power systems. The principle of
operations is to displace the leaking fluid
with the indicating fluid. This can be
done because there is a pressure loss in
the leaking component. When the meter
of Fig. 29 is installed in the hydraulic
power line to the component being
tested, leakage can be measured by the
displacement of the separation level
between the two different liquids in
column A as compared with column B
or C.
In another type of flow meter, leakage
flow in the line between meter and
component under test is measured by the
FIGURE 29. Connection of delta-vee U-tube manometer for
leakage measurements, with tubes (A,B,C) of different
diameters.
Shutoff valves
Regulated pressure
source
Test system
A
B
C
Pressure Change and Flow Rate Techniques for Determining Leakage Rates
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
207
displacement of a bellows. The deflection
of the pressure difference sensing bellows
system varies the setting of a
potentiometer. An output electrical signal
from the potentiometer indicates leakage
directly in volume units. This bellows
system replaces the observation of
movements of a liquid slug in a capillary
tube. Each of the preceding types of flow
meter will work with liquids as well as
with gases, provided that the indicating
liquid slug is immiscible in the fluid
whose leakage is being measured. This
versatility makes the sealed volume leak
testing techniques extremely useful for
leak testing under operational conditions.
FIGURE 30. Mass flow meter with thermal sensor that
measures flow through capillary tube: (a) photograph;
(b) schematic of components of thermal mass flow
transducer; (c) temperature distribution under static (no-flow)
and flowing conditions in flow meter transducer system.
(a)
Fast Response Thermopile
Mass Flow Meter
208
Leak Testing
(b)
Direct current voltage source
Heater
Thermocouple 2
Thermocouple 1
Meter
Heat sink
Heat sink
(c)
Tube temperature (relative units)
The flow meter (Fig. 30a) comprising a
sensor, electronic circuitry and a shunt
measures gas flow rate from 0 to
60 Pa·m3·s–1 (0 to 600 std cm3·s–1). The
shunt causes the flow to divide such that
the flow through the sensor is a precise
percentage of the flow through the shunt.
The circuit board amplifies the sensor
output linearly to a 0 to 5 V direct current
signal proportional to the flow rate.
A thermal sensor measures flow
through a capillary tube. This flow is a
fixed percentage of the total flow through
the instrument. This sensor develops an
essentially linear output signal
proportional to flow, which is about
0.8 mV full scale magnitude (Fig. 30b).
This signal is amplified by the meter
circuitry so that the full scale output is
5.00 V direct current. The output is routed
to interface terminals and to decoding
circuitry in the display.
Measurement of flow rates higher than
60 Pa·m3·s–1 (600 std cm3·s–1) full scale is
achieved by dividing the flow with a fixed
ratio shunting arrangement. The
measuring capillary tube is placed parallel
with one or more dimensionally similar
channels, call laminar flow elements. The
sensor only needs to heat the gas passing
through the capillary tube while retaining
all the mass measuring characteristics.
The fast response of this instrument at
very low rates of air flow permits fast,
accurate leak testing by manual or
automatic means. Table 14 lists
multiplication factors for the air scale
meter indications when this flow meter is
used for gases other than air.
The metal capillary tube of Fig. 30b is
heated uniformly by current from the
transformer. The temperature distribution
is symmetrical about the tube midpoint
with zero flow (Fig. 30b). The external
thermocouples TC1 and TC2 develop equal
but opposing electromotive force outputs
with a symmetrical temperature
Zero flow
Small flow
TC-2
TC-1
L/2
0
L/2
Length of tube (relative units)
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
distribution. When air or gas flows
through the tubing, heat is transferred to
the gas and back again, creating an
asymmetrical temperature distribution
(Fig. 30c). For constant power input to the
tube, the differential thermocouple output
voltage is a function of the mass flow rate
and heat capacity of the gas. Changes in
TABLE 14. Multiplication factors for different gases of
mass flow meter air scale.a
Gas
Acetylene
Air
Ammonia
Argon
Arsine
Bromine
Butane
Butene 1
Carbon dioxide
Carbon monoxide
Chlorine
Chlorine trifluoride
Cyclopropane
Diborane
Ethane
Ethene (ethylene)
Ethylene oxide
Fluorine
Helium
Hydrogen
Hydrogen chloride
Hydrogen fluoride
Hydrogen sulfide
Isobutane
Krypton
Methane
Neon
Nitric oxide
Nitrogen
Nitrous oxide
Oxygen
Pentaborane
n-Pentane
Phosphine
Propane
Refrigerant-11
Refrigerant-12
Refrigerant-13
Refrigerant-14
Refrigerant-22
Refrigerant-114
Silane
Sulfur dioxide
Sulfur hexafluoride
Tungsten hexafluoride
Uranium hexafluoride
Water vapor
Xenon
Conversion Densityc
Symbol
Factor b
(g·L–1)
C2H2
NH3
A
AsH3
Br2
C4H10
C4H8
CO2
CO
Cl2
ClF3
C3H6
B2H6
C2H6
C2H4
C2H4O
F2
He
H2
HCl
HF
H2S
C4H10
Kr
CH4
Ne
NO
N2
N2O
O2
B5H9
C5H12
PH3
C3H8
CCl3F
CCl2F
CClF3
CF4
CHCIF2
CClF2
SiH4
SO2
SF6
WF6
UF6
H2O
Xe
0.67
1.00e
0.77
1.43e
0.76
0.88
0.30
0.34
0.73e
1.00e
0.85
0.45
0.52
0.50
0.56
0.69
0.60
0.93
1.43e
1.03e
1.01
1.00
0.85
0.31
1.39
0.69e
1.38
1.00
1.02e
0.75
0.97e
0.15
0.22
0.79
0.32e
0.36
0.36e
0.42
0.48
0.43e
0.22e
0.68
0.70
0.28
0.23
0.23
0.80
1.37
1.09
1.20
0.71
1.66
3.25
5.98
2.51
2.40
1.84
1.17
2.98
3.78
1.75
1.15
1.26
1.17
1.79
1.58
0.17
0.08
1.48
1.53
1.43
2.48
3.49
0.68
0.84
1.24
1.17
1.85
1.33
2.83
3.18
1.53
1.89
5.93
5.13
4.59
3.65
3.65
6.99
1.33
2.72
6.43
8.22
14.65
0.76
5.54
Relative
Specific
Gravity d
0.90
1.00
0.59
1.38
2.70
4.96
2.08
1.99
1.53
0.97
2.47
3.14
1.45
0.95
1.05
0.97
1.49
1.31
0.14
0.07
1.23
1.27
1.19
2.06
2.90
0.56
0.70
1.03
0.97
1.54
1.10
2.35
2.64
1.27
1.57
4.92
4.26
3.81
3.04
3.03
5.80
1.10
2.26
5.34
6.82
12.16
0.63
4.60
a. No corrections or compensations for temperature or pressure of gas
required.
b. Multiply air scale by these conversion factors.
c. Density in grams per liter at 20 °C (70 °F) and 100 kPa (1 atm).
d. Specific gravity (air = 1.00).
e. Empirical data; other data is theoretical. Example: Flow meter NALL-1K,
0–1000 std cm3·s–1 in air would be 1000 × 1.43 = 1430 std cm3·s–1 at
full scale in helium.
gas composition requires only a simple
multiplier of the air calibration to account
for the differences in heat capacity. The
flow meter can be used for a wide variety
of gases during leakage rate testing. The
full scale flow through the flow meter is
about 1 Pa·m3·s–1 (10 std cm3·s–1).
Figure 31 shows typical arrangements
for leak testing of small items. Figure 31a
shows a pneumatic bridge arrangement.
The object to be tested and an identical,
leaktight part used as a reference volume
are charged with air at pressures up to
135 kPa (20 lbf·in.–2) gage. The effects of
adiabatic heating or cooling of the air
during the pressurizing cycle should be
avoided. The flow meter is then
connected between the unknown and
reference parts to detect any evidence of
leakage (which would allow the pressure
to decrease in the part under test).
Because the adiabatic effects are nearly
identical in the reference and the test
parts, the thermopile flow meter quickly
detects the leakage rate without requiring
a waiting period for attainment of full
equilibrium in temperatures and
pressures. Leakage testing may also be
done by a direct inline leak testing
procedure, as sketched in Fig. 31b, but
this test procedure requires a longer time
cycle than the differential flow
measurement technique.
Orifice Flow Detector with
Differential Pressure
Transducer
Figure 32a shows a leakage test
instrument system that uses an orifice to
convert flow across the orifice element
into a pressure differential sensed by the
differential capacitance sensor (see also
Fig. 14a). The orifice (which produces a
pressure loss when air flows through it) is
connected in series with the air supply
line to the item under test, as shown in
Fig. 32). This system is used with
automatic flow and leakage testers
providing fully automatic cycling and
accept/reject test indicators and output
signals. Leakage sensitivity and
stabilization time are both programmable.
A compensation network provides a
programmed electronic time base signal to
match the dynamic characteristics of
short time cycle flow measurements.
Figure 32a is a photograph showing
instrument connections to the differential
pressure transducer (capacitance gage) and
the orifice or flow restriction element (in
this example, a short length of tubing).
Typical ranges vary from 0.05 to 250 L·s–1
(0.002 to 9.0 ft3·min–1). The dynamic
range is indicated to be 50:1 for operation
with laminar flow pressure loss elements.
Pressure Change and Flow Rate Techniques for Determining Leakage Rates
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
209
Total leakage test cycle times from 0.5 to
2 s are obtainable. The flow compensation
network allows dynamic air flow
measurements without requiring a
stabilization period.
instantaneously in standard engineering
units and in the range selected. A self
check mode provides means for verifying
the integrity of the flow monitor both
electrically and pneumatically.
Digital Electronic Flow
Meter for Monitoring
Leakage Rate Tests
Vacuum Pumping
Technique of Leak Testing
with Flow Measurements
Figure 33 shows a portable, digital
electronic flow meter designed for fast,
accurate indication of leakage rates of
pressurized components such as valves,
O-ring seals, pressure vessels, holding
tanks, tank cars and processing vessels. It
provides a broad range of flow rate
measurements up to 2 × 102 Pa·m3·s–1
(2 × 103 std cm3·s–1) with a resolution of
one part in 2000. Repeatability is
indicated as ±0.2 percent of full scale and
accuracy is ±1.0 percent of full scale
values. The instrument includes a broad
range, high accuracy solid state digital
flow indicator, combined with an integral
flow regulator and a digital pressure
indicator. The only additional equipment
required is a pressure source such as
instrument air or nitrogen for tests as
pressures up to 400 kPa gage (0 to
60 lbf·in.–2 gage) or 700 kPa gate (0 to
100 lbf·in.–2 gage) and means for making
connections to the test unit. Once the
system and object under test have been
pressurized, operation is changed from
the charge mode to the leakage test mode.
The leakage rate indication is displayed
If the test system can be safely evacuated,
leakage can be measured directly by
means of flow meter with vacuum
pumping arrangement sketched in Fig. 34.
The system under test is evacuated
through an opened isolation valve
connected to the vacuum pump inlet. The
exhaust gases from the vacuum pump go
through a surge tank to the flow meter. A
bypassing valve around the pump
provides an alternative path between the
isolation valve and the surge tank. Before
performing the leak test, the vacuum
pumping system leak tightness is first
determined by closing the isolation valve
and measuring the rate of gas flow
through the flow meter. If this flow is
negligible, the isolation valve is then
opened and the flow meter readings are
taken only after an equilibrium (constant
flow rate) condition has been achieved.
The vacuum pressure in the system
under test is adjusted by means of the
bypass valve, which controls the backflow
of gas from, the exhaust port of the
vacuum pump to its inlet port. The lower
limit of vacuum pressure for which the
FIGURE 31. Arrangements for leak testing with thermopile air flow meter: (a) pneumatic bridge leakage testing arrangement
with thermopile flow meter arranged to measure difference in pressure between test object and an identical leaktight object
(reference volume); (b) inline leakage testing arrangement in which test part is pressurized, line valve is closed and leakage is
indicated by pressure drop in flow meter sensing element.
(a)
Bridge arrangement
Regulator
Test part
Valve
Valve
Transducer
Valve
Air source
Reference part
Flow indicator and alarm
Inline arrangement
(b)
Regulator
Valve
Transducer
Test part
Air source
Flow indicator and alarm
210
Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
vacuum pumping leak analysis technique
is useful is in the range of 3 kPa (25 torr).
The lower limit of leak testing sensitivity
is about 0.1 Pa·m3·s–1 (1 std cm3·s–1) and is
mainly dependent on the availability of
suitable flow meters for the vacuum
pressure range used during the leak test.
Sealed Volume Flow Meter
Leak Testing of Nuclear
Containment Systems
Sealed volume leak testing techniques are
also used on large volume systems such as
nuclear containment systems. For this
application, this procedure is commonly
called a verification test. Its purpose is to
verify the accuracy of the leakage test
results and instrumentation used in that
test. It also verifies the validity of the
dewpoint and temperature sensor
locations within the containment
structure. Flow meters used in these large
scale leakage rate tests include thermal
mass flow sensors, rotameters and
integrating gas flow meters usually with
ranges of 25 to 700 Pa·m3·s–1 (0.5 to
15 std ft3·min–1). These flow meters are
usually designed for the planned leak
testing conditions and they produce
readouts compensated to standard
pressure and temperature conditions. The
accuracy of the flow meter must be
FIGURE 33. Portable digital electronic flow meter for
monitoring leakage rates in pressurized systems.
FIGURE 32. Air flow meter with orifice and differential
pressure transducer: (a) photograph; (b) pneumatic circuit.
(a)
FIGURE 34. Arrangement for vacuum pumping technique of
leakage measurement with flow meter.
(b)
Dial
gage
Quick
disconnect
Pressure
transducer
Air
supply
Out
Pressure
regulator
In
System
under
test
Isolation
valve
Bypass
valve
Test
item
Surge
tank
Solenoid
valves
Pump
Quick
disconnect
Flow meter
Pressure Change and Flow Rate Techniques for Determining Leakage Rates
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
211
commensurate with the accuracy of the
leakage rate test instrumentation and also
with the accuracy required in the
containment leakage rate test results.
Procedures for Flow Meter
Verification Test of Nuclear
Containment Systems
The verification test is normally
performed as the last phase of a
containment test. It follows the test for
the system measured leakage rate Qam
(usually given as a percentage of air mass
lost in 24 h). The flow meter is installed
in the system with a valve to isolate it
from the system under test. The
verification test may be performed by
measuring either the out-leakage or the
in-leakage that passes through the flow
meter. For either technique, a meter valve
is placed downstream from the direction
of leakage flow through the flow meter, to
minimize the pressure loss across the flow
meter. After opening the isolation valve
between the test system and the flow
meter, this metering valve is adjusted to
produce a leakage flow through the flow
meter from (or into) the test system that
is some required percentage (usually 75 to
125 percent) of the allowable leakage rate
Qa for the system under test.
The leakage rate test of the
containment is then continued. After a
period of 4 to 6 h with a minimum of ten
sets of data, the combined leakage rate Qc
of the containment system and flow
meter and the leakage rate Q0 of the flow
meter are determined using the flow
meter readings. The difference between
these two leakage rates is Qc – Q0 = Q´am.
This difference Q´am in reading is then
compared to the leakage rate Qam
measured previously on the containment
test system alone, before the inflow or
outflow of air from the containment
through the flow meter. The two values
must agree with 25 percent of the
measured containment leakage rate Qam.
This is to say that Qam – Qam must be
equal to or smaller than 0.25 Qam.
True Thermal Mass Flow
Meters for Accurate Flow
Rate Measurements
The containment verification test just
described requires a true mass flow meter
that measures the mass of gas that passes
through it. Figure 35 shows a true mass
flow sensor element which does not
require temperature or pressure
compensation and provides ±1 percent of
full scale accuracy and linearity. The
sensor unit has a stainless steel flow tube.
A heater coil is wound around the center
212
Leak Testing
section of its length. Sensor coils are
wound around the flow tube on either
side of the heater coil and are connected
in a bridge circuit. The zero flow, the
bridge circuit is balanced and the output
signal is zero. With flow the sensor coils
detect the resulting temperature
difference, which is proportional to mass
flow. The output electrical signal varies
linearly with the gas flow rate. Signals can
be used for measuring, recording or
controlling gas flow rates with valves and
an automatic controller. Sensors for
specific gases such as air, nitrogen,
hydrogen, oxygen and helium are
FIGURE 35. Thermal mass flowmeter uses a true mass flow
sensor for measuring gas flow rates accurately: (a) sensor;
(b) principle of operation.
(a)
(b)
To power
supply
Downstream
temperature
sensor
Upstream
temperature
sensor
Bypass sensor
tube
Flow
15 to 28 V
direct current
Bridge
for ∆T
detection
Amplifier
0 to 5 V direct
current and
4 to 20 mA
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
available with ranges from 0 to 0.015 up
to 0 to 10 Pa·m3·s–1 (0 to 10 up to 0 to
5 × 103 std cm3·min–1). Repeatability of
indications is claimed as ±0.2 percent of
full scale. Output signals from the thermal
mass flow sensors of Fig. 35 can actuate
indicating meters or provide 0 to 5 V
direct current signals that can be
transmitted up to 300 m (1000 ft) to
recording instruments, digital indicators
or controllers. The electrical output signal
is linearly proportional to the mass flow
rate through the sensor.
Flow Meter Tests to Locate
Leaks in Gas Filled Electric
Power Cables
Electric utility companies have made use
of a U-tube manometer equipped with
appropriate valving to use as a flow meter
for locating gas leaks in gas pressurized
electric power cable sheaths. When the
manometer is installed in a segment of
the pressurized gas filled cable sheathing,
oil will rise in the glass tube of the
manometer, on the side closer to the leak.
In this test, the manometer measures the
pressure loss in the segment of cable
across which it is connected, when gas
flows toward the leak.
Pressure Change and Flow Rate Techniques for Determining Leakage Rates
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
213
References
1. CRC Handbook of Chemistry and
Physics. Cleveland, OH: Chemical
Rubber Company (1964).
2. Fleshood, D.L. “Containment Leak
Rate Testing: Why the Mass-Plot
Analysis Method Is Preferred.” Power
Engineering. Barrington, IL: Technical
Publishing Company (February 1976):
p 56-59.
3. Lau, L.W. “Data Analysis during
Containment Leak Rate Test.” Power
Engineering. Barrington, IL: Technical
Publishing Company (February 1978):
p 46-49.
4. Kendall, M.G. and A. Stuart. The
Advanced Theory of Statistics, third
edition. Vol. 2. New York, NY: Hafner
Publishing Company: p 130-132.
5. Tietjen, G.L., R.H. Moore and R.J.
Beckman. “Testing for a Single Outlier
in Simple Linear Regression.”
Technometrics. Vol. 15, No. 4.
Alexandria, VA: American Statistical
Association (November 1973):
p 717-721.
6. ANSI/ANS-56.8-1981, Containment
System Leakage Testing Requirements,
Appendix C. La Grange Park, IL:
American Nuclear Society (1981).
7. Guthrie, A. Vacuum Technology. New
York, NY: John Wiley and Sons (1963).
Reprint, Malabar, FL: Krieger
Publishing (1990).
8. Steinherz, H.A. Handbook of High
Vacuum Engineering. New York, NY:
Reinhold Publishing Corporation
(1963).
214
Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
C
6
H A P T E R
Leak Testing of Vacuum
Systems
Charles N. Sherlock, Willis, Texas
Carl A. Waterstrat, Varian Vacuum, Lexington, Massachusetts
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
PART 1. The Nature of Vacuum
Definition of a Vacuum
The word vacuum is derived from the
Greek word meaning empty. In practice,
use is made of some type of vessel
(vacuum enclosure, chamber or container)
to contain a vacuum. When the enclosure
is closed to the surrounding atmosphere
and air or gas is removed by some
pumping means, a vacuum is obtained.
Various degrees of vacuum can be
obtained, depending on how much air is
removed from the enclosure. Common
terms such as partial vacuum, rough
vacuum, high vacuum and ultrahigh
vacuum are used to describe degrees of
vacuum. A vacuum is any pressure below
the prevailing atmospheric pressure.
Practically speaking, a vacuum such that
the containing vessel is empty, i.e., free of
all matter (molecules), is never obtained.
If this were possible, the vacuum would
be called a perfect or absolute vacuum.
Applications of Vacuum
Environments
Vacuum is used to reduce the interaction
of gases or air with solids and to provide
control over electrons and ions by
reducing the probability of collision with
molecules of air. Vacuum pumps are used
by industry and laboratories to create a
vacuum environment for these
operations. Most gases react with solids to
cause effects such as oxidation, which it
may be necessary to avoid. In a vacuum
environment, the necessary operation
may be performed so that undesirable
effects are reduced or eliminated. For
example, unless most of the air is
removed from an incandescent light bulb,
oxygen in its atmosphere will react with
the hot tungsten filament, causing it to
burn out prematurely. An electron tube
could not operate at atmospheric pressure.
Electron flow would be impeded by
collision with air molecules due to the
extremely small mean free path. In
addition, elements within the tube may
react with the air. Other examples can be
cited where vacuum is necessary to
produce desired results that could be
unattainable in any other way.
Vacuum is required in many industries
and products. In addition to light bulbs
216
Leak Testing
and computer chip manufacturing,
vacuum is used in magnetrons, cathode
ray tubes, television picture tubes,
semiconductor devices, solar cells, plating
metals and plastics, thin film deposition,
lifting objects, plasma physics, cryogenics,
metallurgical processing, electron beam
welding, brazing, distillation organic
chemistry, packaging, mass spectrometry,
space simulation and leak detection.
Many other areas find application for
vacuum equipment.
Changes in Pressure Units
Used for Vacuum
Measurements
The presently preferred SI unit for
pressure is the pascal (Pa). The standard
atmospheric pressure at sea level and 0 °C
(32 °F) is equal to 101.325 kPa. Earlier
units used for pressure in vacuum relate
to atmospheric pressure indicated by the
height (nearly 760 mm) of the mercury
barometer column at sea level and 0 °C
(32 °F). The unit known as the torr was
defined as 1/760th of the pressure of the
mercury column. The torr was named in
honor of an Italian physicist, Evangelista
Torricelli (1608-1647), inventor of the
mercury barometer. The torr is almost
identical to the millimeter of mercury
(mm Hg), because there are 759.96 torr in
a standard atmosphere. The difference
between the two units amounts to so little
that torr and mm Hg have been used
interchangeably.
Variation of Atmospheric
Pressure with Altitude
The mercury barometer is a device for
measuring atmospheric pressure. As the
altitude increases, the pressure decreases
because fewer gas molecules press on any
surface. A knowledge of how the pressure
changes with altitude is very important in
connection with various space studies.
Table 1 shows the relationship between
pressure and altitude in the earth’s
atmosphere.
At an altitude of 50 km (27 mi) the
pressure is about 0.1 percent of standard
atmospheric pressure or 100 Pa
(0.015 lbf·in.–2). Air at this altitude
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
contains one thousandth of the number
of molecules per unit volume in air at sea
level. At 400 km (250 mi) altitude, the
pressure is in the range of 1 µPa or 10–11
parts of sea level pressure. Table 2 gives a
relative measure of gas characteristics at sea
level and at 1 nPa (10 ptorr). Compared
with the number of molecules in a cubic
centimeter at atmospheric pressure, it is
seen that there are one hundred
thousandth of one millionth as many
molecules at 10–6 Pa (1.5 × 10–10 lbf·in.–2).
However, a tremendous number of
molecules (3 × 108) still remain in a cubic
centimeter at a pressure of 1 µPa (10 ntorr).
Pressures around 1 µPa (10 ntorr) are not
uncommon in good vacuum systems.
Diffusion and Adsorption
of Gases in Vacuum
Systems
The kinetic molecular theory of gases and
the ideal gas laws (Boyle’s, Charles’,
Dalton’s and the general gas law), are
applicable to vacuums. In vacuum, fewer
molecules are dealt with, but their basic
behavior is predictable by the molecular
theory of gases and does not change. The
TABLE 1. Change in atmospheric pressure with altitude.
Altitude
_______________
km
(mi)
0
1
2
5
10
(0.6)
(1.2)
(3.1)
(6.2)
20
50
100
200
500
1000
(12.4)
(31.1)
(62)
(124)
(311)
(621)
ability of a gas to diffuse increases when
its pressure is reduced. Consider the
example of ammonia vapor being released
in a room. The reason that it is not
detected immediately at the other end of
the room is that the path each ammonia
molecule takes is restricted by the air
molecules with which it collides. It is only
after many billions of collisions with air
molecules that the ammonia molecules
finally make their way across the room. If
the room were pumped down a high
vacuum, there would be many fewer air
molecules and far fewer collisions to
impede the path of the ammonia
molecules. Thus, an ammonia molecule in
a high vacuum takes less time to complete
its trip across a given distance than in
gases at higher pressures.
Only those molecules that are in
motion within a vacuum chamber create a
pressure through collisions with its walls.
A molecule that is adsorbed to the wall
surface is stationary and does not produce
collisions. Therefore, adsorbed gas
molecules do not contribute to the total
pressure. However, molecules adsorbed on
surfaces can be returned to the gas phase
by thermal agitation produced by the
application of heat. Thus, outgassing
effects can contribute to pressure in an
evacuated system, because a molecule can
undergo repeated collisions and exert
pressure only when it is in the gaseous
state.
Pressure
_________________________
kPa
(atm)
101.325
(1.00)
89.90
79.50
54.00
26.50
(0.887)
(0.785)
(0.533)
(0.262)
5.53
7.98 × 10–2
3.2 × 10–5
8.5 × 10–8
3.0 × 10–10
7.5 × 10–12
(0.055)
(7.9 × 10–4)
(3.2 × 10–7)
(8.39 × 10–10)
(3 × 10–12)
(7.4 × 10–14)
Remarks
international
standard
jetliner
altitude
low orbit
Mean Free Path of Gases
in Vacuum Systems
At normal atmospheric pressure, gas
molecules make many collisions with
each other. The average distance that a
molecule travels before colliding with
another molecule is known as the mean
free path. The mean free path of two
different gases at the same pressure will
not be the same; this is because the mean
free path depends on the molecular size,
which varies from one gas to another. In
spite of this fact, it is still possible to give
a useful relationship between mean free
path and pressure. The approximate
values of mean free paths for air and
TABLE 2. Comparison of atmospheric properties at sea level and at high altitude.
Condition
Pressure
Number of molecules in 1 cm3 (0.06 in.3)
Mean free path
Time to form a monolayer of adsorbed gas on a clean surface
Average speed of nitrogen molecule at room temperature 20 °C (68 °F)
At Sea Level
At 400 km (250 mi) Altitude
101.325 kPa (760 torr)
2.7 × 1019
93 nm (3.7 × 10–6 in.)
>10 ns
1600 km·h–1 (1000 mi·h–1)
1 µPa (10 ntorr)
3 × 108
9.3 km (5.8 mi)
120 s
1600 km·h–1 (1000 mi·h–1)
Leak Testing of Vacuum Systems
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
217
other gases are given as a function of gas
pressure in Eq. 1:
(1)
λ
=
0.0095
P
where λ is mean free path (meter) and P is
gas absolute pressure (pascal).
Effects of Molecular Friction and
Gas Viscosity in Viscous Flow
As shown by Eq. 1, the mean free path
length varies inversely with absolute
gaseous pressure. The concept of mean
free path is useful in describing vacuum
ranges. The mean free path at
atmospheric pressure is very short (see
Table 2), due to the large molecular
density. Therefore, collisions occur much
more frequently between gas molecules
than between molecules and the walls of
the container. Thus, the gas acts much
like a fluid. Under a pressure differential
this gaseous fluid moves as a unit and is
considered to flow. The molecules, while
drifting slowly in the direction of flow,
move rapidly along random paths. Any
resistance to this flow is due to the
viscous properties of the gas. The term
viscous refers to molecular friction and is
used to describe the flow characteristics of
a fluid. Water, for example, is less viscous
than syrup because it flows or pours more
easily. The cross sectional dimension of
the container or tube through which the
gas flows is important because it
determines the velocity of the molecules
within the flowing gas. The viscous
properties of the gas are functions of the
gas viscosity and the gas velocity. When
viscous properties control gas flow rates,
the situation is termed viscous flow.
Effects of Mean Free Path and
Flow Cross Section on Molecular
Flow of Gas
As the pressure of the gas within a system
is reduced, the mean free path of the
molecules increases and the flow
characteristics change gradually. As the
mean free path becomes comparable to
the cross sectional dimensions of the
tube, collisions occur less frequently
between molecules and the apparent
viscosity of the gas decreases. Under these
conditions, the event that is most likely
to affect the direction of the molecules;
travel is a molecular collision with the
tube wall. As the pressure is further
reduced, the mean free path becomes
greater than the tube’s cross sectional
dimensions. The diameter of the tube
alone then determines the resistance to
flow; this situation is called molecular
218
Leak Testing
flow. The motion of a particular molecule
is entirely random and unpredictable; it is
as likely to move in one direction as in
any other direction. To a molecule, tube
wall appear very rough and irregular. The
direction of molecule rebound after
impact with the tube wall thus tends to
be independent of the direction of
incidence. (This is an over simplified
description.) Figure 1 is a sketch of
particle motions during molecular flow of
gases through a tube. Note that not all of
the molecules entering at the left exit at
the right. The gas flow will continue as a
net movement to the right only as long as
there is some driving force causing
movement from left to right. As gas
concentration gradient is such a force. A
pressure differential is another force that
can control the direction of net flow of a
gas. Both can contribute to flow of a
tracer gas through a leak.
Specifying Gas Flow Rates
The flow rate of liquids is expressed
simply as so many volume units per unit
time, such as liters per second. When,
however, the flow rate of gases is
considered, it is necessary to know not
only the volume of a gas but its pressure
and temperature as well. A cubic meter
volume of gas at 100 kPa (15 lbf·in.–2)
pressure and a temperature of 20 °C
(68 °F) will contain ten times as many
molecules as a cubic meter volume of gas
at 10 kPa (1.5 lbf·in.–2) and 20 °C (68 °F).
Only a complete statement of volume,
displacement rate, gas pressure and
temperature can accurately describe the
total quantity of gas that flows per unit of
time. In both liquids and gases, it is mass
flow that is of interest. For liquids of
constant density, the mass rate of flow is
directly proportional to volume flow rate.
With gases, density varies both with
temperature and with pressure. Thus, for a
given gas, volume displacement rate,
pressure and temperature must be known
to define the mass flow rate.
FIGURE 1. Molecular motion along a tube, with particle mean
free path far larger than tube diameter.
3
1
2
5
4
2
5
4
3
1
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
The Concepts of Gas
Quantity and Pumping
Speed
From the gas laws, it is known that the
product PV of pressure P and volume V is
proportional to the number of molecules
in a sample of gas. In static systems, the
PV product is constant at a given
temperature. This product PV is known as
the quantity of gas. Common units of gas
quantity include torr liter (torr-L); the
atmospheric cubic centimeter (cm3 of
volume at standard sea level atmospheric
pressure or std cm3); and the bar liter
(bar-L). The preferred SI unit of gas
quantity is the pascal cubic meter (Pa·m3).
In steady flow, the same quantity of
gas (number of molecules) that enters one
end of a tube must leave at the other end,
even though there may be different
volumes of gas entering and leaving per
unit time. If the PV product is used as a
measure of the amount of gas flowing
through a tube, computation may be
done with a minimum of complication.
The volumetric pumping speed S is the
time rate of volume displacement, as
given by Eq. 2:
(2)
S
=
V
t
Typical units of pumping speed S
would be cubic meter per minute
(m3·min–1), cubic meter per second
(m3·s–1), liter per second (L·s–1) and cubic
foot per second (ft3·s–1).
Concepts of Throughput
and Leakage Rate
In vacuum practice, the preferred
description of the rate of flow of gas is
commonly called throughput.
Throughput is the quantity of gas or a
measure of the total number of molecules
at a specified temperature, passing an
open section of the vacuum system per
unit time. Leakage rate is a similar
measure of the total number of molecules
at a specified temperature passing through
a leak per unit time. Q is the symbol
commonly used for gas throughput per
unit time, in pascal cubic centimeter per
second:
(3)
Q
=
PV
t
By combining Eqs. 2 and 3, the product of
pumping speed S and gas pressure P can
be equated to throughput by Eq. 4:
(4)
Q
=
S × P
Equation 4 is the universal relationship
on which vacuum pumping throughput
calculations are based. As an example of
its use, suppose the gas in the pipe
between Sections 1 and 2 of Fig. 2 passes
Section 1 in 1 s and this volume V is
100 L (0.1 m3) and pressure P at Section 1
is 10–4 Pa and displaced volume V =
0.1 m3, divided by the time t = 1 s:
Q
=
S × P
=
=
10 −4 × 0.1
PV
t
= 10 −5
Comparison of Gas Flow with
Liquid Flow
Before attempting a more thorough
discussion of gas flow, it may be helpful
to compare gas flow with water flow. To
get any fluid to flow within a pipe, a
pressure differential must be established
between the two sections across which
the fluid is to flow. (Gravitational effects
are neglected in this introductory
discussion.) The fluid would then flow
from the high pressure region P1 to the
low pressure region P2. Consider a closed
system of pipes through which water is
circulated as in an automobile. The water
pump creates the pressure differential
necessary for water to flow. Across each
component (radiator, engine block,
thermostat, different sizes of piping) in
the system, the pressure drops. The sum
of all these pressure drops equals the
pressure differential across the water
pump. The magnitude of the pressure
drop across each component of the
system depends on its physical geometry.
Clearly, a smaller diameter pipe will result
in decreased flow for the same size pump.
similarly, increasing the length of the pipe
will reduce the flow, whereas decreasing
the length of the pipe will increase the
flow. Shorter lengths and larger diameters
reduce the resistance to flow through the
pipe.
Analogy of Gas Flow to Electric
Current through Resistance
The analogy may be carried further by
comparing the gas flow system to an
electrical circuit. Given an electrical
circuit with a battery and a resistor in
series with it, the battery may be
considered to be the pump and the
resistor the pipe. Increasing or decreasing
the resistance decreases or increases the
current flow (analogous to gas flow),
Leak Testing of Vacuum Systems
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
219
respectively. If the circuit consists of a
series of resistors and a battery, the sum of
the voltage drops across each of the
resistors (pressure drops) is equal to the
total voltage generated by the battery
(pressure differential created by the
pump). The voltage drop across each
resistor will depend on the magnitude of
resistance of that component. The larger
the resistance, the larger the voltage drop
(see Fig. 3).
Gas Conductance and Its Electrical
Analogy
In vacuum, one speaks not of the
resistance a tube or component offers to
gas flow, but instead uses the reciprocal
term conductance. Conductance is a
measure of the ability of a vacuum
component to permit gas flow or not to
impede it. Consequently, the greater the
resistance, the smaller the conductance
and vice versa. Figure 3 shows the
electrical analogy of a tube in a vacuum
system. The battery is analogous to the
vacuum pump, current is analogous to gas
flow and the resistor is analogous to pipe.
FIGURE 2. Rate of flow of a gas Q through tube with applied
pressure differential ∆P = P1 – P2 (P1>P2). (See analogous
electric circuit of Fig. 3.)
Section 2
Section 1
Tube conductance = C
Tube resistance = R
P2
(5) ∆ P
=
P1 – P2 = Q × R =
Equation 6 is the defining equation for
gas conductance: the ratio of throughput
Q to pressure differential ∆P across the
conductance.
Gas Conductance with
Sequential Tubes of
Passages
If two different diameter pipes with
different conductance values are
connected in series as in Fig. 4a, the total
conductance of the connection between
extreme ends decreases (resistance
increases). From Eq. 6, the conductance of
the pipe between Sections 1 and 3 may be
expressed as in Eq. 7:
Q
P1 − P3
=
From vacuum
chamber
Gas flow rate
Gas flow rate Q = (P1 – P2)C
Q
C
Because R is equal to 1/C, Eq. 5 may be
written in the form of Eq. 6 for gas
conductance C:
Q
C =
(6)
∆P
(7) C13
P1
To vacuum
pump
In an electrical circuit, the voltage drop
across a resistor is the product of the
current and resistance. In a vacuum
circuit, the pressure differential across a
pipe is the product of throughput (gas
flow) Q and resistance R. Equation 5 states
this relation mathematically for the
pressure differential ∆P:
(P1
(8) P1 − P3
=
(9) P1 − P2
=
− P2 ) + ( P2 − P3 )
Q
C12
or
Pressure differential ∆P = P1 – P2 = QR = Q/C
P2 − P3
FIGURE 3. Electrical analogy of vacuum pumping pipe
conductance system of Fig. 2. G is the electrical
conductance, the reciprocal of electrical resistance R, so that
G = 1/R. ER = Eb – EL = E1 – E2 = IR = I/G.
=
Q
C 23
Now, by combining Eqs. 7, 8 and 9, the
relationship for C13 becomes:
(10) C13
=
Q
Q
C12
+
Q
C 23
R = 1/G
or, in its reciprocal form:
Eb
i
–
EL
+
220
Leak Testing
Load
(11)
1
C13
=
Q
Q
+
C12
C 23
Q
=
1
C12
+
1
C 23
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
In its general form, Eq. 11 may be written
as Eq. 12:
(12)
I
CT
=
1
1
+
C1
C2
+
1
1
+…+
C3
Cn
(14) Q A
=
=
CA
×
∆P
=
Cb
×
∆P
and
In Eq. 12, the subscript T denotes the
total conductance of a number of
conductances C1, C2, C3 ... Cn connected
in series.
In the case of only two conductances
connected in series (Fig. 4b), Eq. 12
should be written in the form of Eq. 13:
(13) CT
division is not equal but depends on the
conductance of each component. From
Eq. 6, the gas load in each parallel pipe
may be written in the form of Eq. 14:
Qb
The total conductance between points 1
and 2 is
Qa + Q b
(15) C12 =
∆P
Substituting from Eq. 14, Eq. 15 gives:
Ca ∆ P + C b ∆ P
(16) C12 =
∆P
C1 × C 2
C1 + C 2
This case is analogous to the special case
of two electrical resistors connected in
parallel.
Simplifying, Eq. 16 becomes:
Gas Conductance for Pipes
or Tubes Connected in
Parallel
In its general form, the total conductance
for a number of pipes connected in
parallel is equal to the sum of the
individual conductances, as given by
Eq. 18:
Figure 5 shows two lengths of pipe
connected in parallel. In this connection,
the total gas load (throughput) flowing
from the vacuum chamber divides
between the two pipes as shown. The
FIGURE 4. Electrical relationship of two conductances in
series: (a) pipe conductances in series; (b) connection of two
electrical resistances.
(17) C12
(18) CT
=
Ca
=
+
Cb
C1 + C 2 + C 3 + … + C n
FIGURE 5. Electrical analogy of two gas conductances in
parallel: (a) connection of two parallel gas conductances;
(b) electrical circuit analogous to two gas conductances in
parallel.
(a)
P1
P2
Ca
(a)
P3
To
turbomolecular
or diffusion
pump
P2
P1
Qa
Pipe 1
To turbomolecular
or diffusion pump
C12
C23
From
vacuum
chamber
Cb
Q
Q
Pipe 2
Pipe 1
Vacuum
chamber
QT = Qa + Qb
Pipe 2
Qb
1
1
1
— =— +—
C13 C12 C23
(b)
(b)
R1
R1
R2
I1
I
–
EB
(pump)
+
EL
Load
(vacuum
chamber)
R2
–
EB
(pump)
+
EB
I2
EL
Load
(vacuum
chamber)
Leak Testing of Vacuum Systems
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
221
Pumping Speed
In operating a vacuum system, there is an
interest in how fast gases are removed
from the system. The rate of removal of
gases is measured by pumping speed S.
From Eq. 4, pumping speed is defined as
the ratio of the throughput Q to the
pressure P at the point in the system.
Mathematically, this relation is given by
Eq. 19 (m3·s–1):
(19)
S
∆P
=
Pc
− Pp
=
Q
C
FIGURE 6. Net pumping speed relationship applicable to
conductance C between vacuum pump and chamber being
evacuated. Pressure at vacuum chamber (inlet to
conductance C ) is Pc , and pumping speed Sc = Q/Pc.
Pressure at vacuum pump (outlet of conductance C) is Pp
and pumping speed is Sp = Q/Pp. Because vacuum pump
pressure Pp is lower than chamber pressure Pc , whereas Q is
the same at each end of conductance C, the pumping speed
is different at inlet and outlet of C.
C
Diffusion
pump
Q
Sp = —
Pp
222
Leak Testing
Q
=
S p Pp
and pressure as:
Pp
In Eq. 20, the subscripts c and p refer to
the chamber and pump, respectively. The
Sn
— = Qn
Pn
(21) Q p
Q
P
=
If the inlet to a vacuum pump were
connected directly to a vacuum vessel,
then the pumping speed at the vessel
would be the same as that at the pump
inlet. Because it is physically impossible to
join the pump and vessel without
introducing a connector the pumping
speed at the vessel will be lower than that
at the pump. Pumping speed loss depends
on the magnitude of the conductance
that causes a loss in pressure or creates
differential pressure between pumps and
vessel. Figure 6 is used to help establish a
relationship between the net pumping
speed at the vacuum chamber, pumping
speed at the port of a vacuum pump and
the conductance between them. Although
the connection is shown as a pipe in
Fig. 6, it could be a combination of any
number of vacuum components, each
contributing a value of conductance. The
flow of gas is from the chamber to the
pump. From Eq. 5, the pressure drop is
given by Eq. 20:
(20)
throughput Q is the product of the speed
S and pressure P where each is measured
at the same point, such as at the pump or
chamber. Throughput at the pump is
therefore expressed as Eq. 21:
Q
=
Sp
At the chamber being evacuated,
throughput is expressed as:
(22) Q c
=
Sc Pc
and pressure as:
Q
=
Pc
Sc
Substituting Eqs. 21 and 22 into Eq. 20
results in the relation of Eq. 23:
(23)
Q
Sc
–
Q
Sp
Q
C
=
Rearranging terms:
(24)
Q
Sc
=
Q
Sp
+
Q
C
and multiplying by 1/Q:
(25)
1
Sc
=
1
Sp
+
1
C
In the general case, the net speed Sn at
any point in a vacuum system is related
to the pump speed Sp and the total
conductance Ct between that system point
and the vacuum pump by Eq. 26:
(26)
1
Sn
=
1
Sp
+
1
Ct
Analysis of Eq. 25 shows that, except for
the case of an infinite conductance (zero
resistance), the net speed will always be
less than the pump speed. How much less
depends on the value of the tube
conductance.
Vacuum
chamber
Sc
Qc = —
Pc
Q
Sc = —
Pc
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
PART 2. Principles of Operation of Vacuum
Systems and Components
Introduction to Vacuum
Pumping
To attain vacuum in a container, some
means of pumping must be used. A pump
cannot reach into the system and extract
molecules, but must simply wait for
molecules to wander through the natural
exit in the container and into the pump
for isolation and discharge. When the
pressure in the vacuum system becomes
so low that there is no longer a net
movement of molecules into the pump,
the base pressure of the system has been
reached under those conditions. How low
this ultimate pressure is will be
determined by conditions such as (1) the
leak tightness of the vacuum system,
(2) the nature and condition of materials
within the vacuum system that might
cause outgassing and (3) the operating
characteristics of the pumps in
combination with the specific vacuum
system.
Molecular Conditions
Limiting Rates of Pumping
of Vacuum Systems
To evacuate a closed system initially at
atmospheric pressure, numerous gaseous
molecules must be removed from within
the closed system. The fewer the gas
molecules remaining, the lower the
absolute pressure of gas within the
system. However, the common concept
that vacuum pumps draw out the air like
a vacuum cleaner is wrong. Molecules in
the gaseous phase are in constant motion
and collide with each other and with the
walls of the container. A certain number R
of molecules strike each unit area of the
container wall per unit time. The number
of gaseous molecules striking a unit area
(square meter) of the container wall per
unit time (second) is given by Eq. 27:
(27)
R
=
(2.63 × 10 )
24
P
MT
where R is rate of molecular impact with
the wall in molecules per square meter per
second; P is absolute pressure in evacuated
chamber (pascal); M is the molecular
weight of gaseous particles, in unified
atomic mass unit (u); and T is the
absolute temperature of gas within the
evacuated container (kelvin). A similar
expression can be given for a mass of gas
(kilogram) striking a unit area (square
meter) of the evacuated container wall per
unit time (second):
P
R ’ = 43.8 × 10 −4
MT
(
)
where R is rate of gas mass impact with
wall (kg·m–2.s–2); P is absolute pressure
within evacuated chamber (pascal); M is
molecular weight of gaseous particles, in
unified atomic mass unit (u); and T is
absolute temperature of gas within
evacuated container (kelvin).
If a hole of unit area were cut through
the container wall, those gas molecules
that would have collided with the
container wall in the area of the hole and
rebounded within the container will now
pass through the hole at the rate given by
Eq. 27a. If these escaping gas molecules
are now prevented from reentering the
container through that hole, the net effect
would be that of reducing the number of
molecules within the container and thus
reducing the internal gas pressure. This is
the basic concept of vacuum pumping,
namely, to provide a natural exit for gas
molecules and to isolate the escaping gas
molecules so that they cannot reenter the
container being evacuated. The vacuum
pump cannot extract gas molecules from
within the evacuated container; it merely
aids those molecules that pass through
the hole in the wall to naturally escape
being reinjected into the vacuum
container through that same hole. It is
impossible to pump any gas out of an
evacuated container at any rate faster
than that at which internal gas molecules
strike the hole area by their random
kinetic motions.
Example of Limitations on Vacuum
Pumping of Gaseous Nitrogen
For gaseous nitrogen with a molecular
weight M = 28 at room temperature,
298 K (25 °C or 77 °F) and atmospheric
pressure, 101 kPa (1 atm), Eq. 27a
indicates that the number of molecules
striking each square meter of container
wall during each second would be
calculated as:
Leak Testing of Vacuum Systems
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
223
R
=
=
(2.63
× 10 24
)
× 101 000
28 × 298
2.92 × 10 27
Similarly, if the mass (kilogram) of
nitrogen gas striking the unit area (square
meter) per second were to be determined,
Eq. 27b indicates that this mass would be:
R’
=
×
=
(43.8 × 10 )
−4
×
101 000
28
298
136
The volume flow rate corresponding to
the above mass flow rate would be equal
to 136 × 0.8714 = 118 m3·s–1·m–2
(7.08 × 103 ft3·min–1·ft2).
For the case of molecular nitrogen (N2)
at 298 K (77 °F) and atmospheric pressure,
the conversion factor is 0.8714 m3·kg–1.
Thus, for gaseous nitrogen molecules
(each of which contains two nitrogen
atoms), the volumetric rate of exit of gas
through a hole in the evacuated container
would be 118 m3·s–1·m2 or 11.8 L·s–1·cm2
(7.08 × 103 ft3·min–1·ft2). This is the
maximum rate at which nitrogen can be
pumped from a container whose internal
pressure was 101 kPa (1 atm). As system
pressure decreases, the molecular or mass
rate of pumping drops proportionally.
Conditions Limiting Rate
of Pressure Reduction by
Pumping
Pumping times greater than expected for
reduction of pressure to desired levels can
result from system contamination or
system leaks. System contamination can
be caused by processing of so-called dirty
work materials or by allowing excessive
time without thorough cleaning of the
vacuum equipment. Contamination of
this type results in many layers of various
compounds, organic or otherwise, which
build up on interior surfaces. The
contaminated surfaces then outgas at such
rates that the pump capacity may be
unable to reduce pressure to desired levels
within acceptable pumping times. Water
vapor adsorbed to chamber walls is a
common contaminant. Dirty walls are
subject to more severe water adsorption. It
is also possible for mechanical pump oil
to become contaminated, which alone
can cause poor pumping characteristics.
If pumping is slowed by system leaks,
thorough mass spectrometer leak
detection tests inspection should be
performed and leaks repaired.
224
Leak Testing
Operation of Mechanical
Pumps for Vacuum
Systems
The mechanical pump is an essential
component used in vacuum systems to
evacuate a chamber from atmospheric
pressure to about 0.1 Pa (10–3 torr)
absolute pressure. Of the various types of
mechanical pumps, the rotary oil sealed
vacuum pump shown in Fig. 7 is most
common. The pump consists of a
stationary housing, an eccentrically
mounted rotor with two spring loaded
vanes, an inlet port and a discharge port.
Air enters the pump from the vacuum
chamber through the inlet port. This air is
trapped, compressed and ejected into the
atmosphere through the discharge port by
means of the rotor arrangement. Sealing
of the eccentric rotor vacuum pump is
done by an oil film between the two
sliding spring loaded vanes that make
contact between the rotor and the
housing. Oil is used as the pump sealant.
Close tolerances must be maintained to
prevent leaks and by passing of gases.
Consequently, care must be taken to
prevent solid particles from entering the
pump. Each rotation of the rotor
discharges two volumes; each volume is a
certain percentage of the volume to be
evacuated. This would indicate that even
a perfect pump could never evacuate to a
vacuum linearly but could only approach
this condition as an exponential function
of pumping time.
FIGURE 7. Rotary mechanical vacuum pump with eccentric
rotor and spring loaded vanes. Pump oil provides a sealing
film at points of vane contact with stator housing.
Outlet
Inlet
Rotor
Vane
Oil
Spring
Housing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
At low chamber pressures, air may leak
back into the evacuated volume. This can
be minimized by putting two pumps in
series so that the discharge from the first
pump chamber is not directly to
atmosphere but to some intermediate
pressure maintained by the second (or
backing) pump.
Pump Oil Used in Rotary Vacuum
Pumps
The operating fluid in any type of pump
is called the pump fluid or pump oil.
With rotary pumps, normally a good
quality light petroleum oil, with the high
vapor pressure factions removed, is used
to provide pump sealing and lubrication
between the rotor vanes and stator
housing. The oil for lubricating and
sealing is contained in an oil reservoir.
The arrangement of the reservoir differs
from manufacturer to manufacturer. In
some small pumps, the pump chamber is
actually immersed in the reservoir,
whereas for the larger pumps the reservoir
is usually separated from the pump
chamber, often being mounted above the
pump itself.
Prevention of Condensate
Contamination of Pump
Oils
Contamination of pump oil is one of the
main difficulties with rotary pumps. As
the gases and vapors are compressed, the
vapors will tend to condense and
contaminate the oil. Degassing of vapors
from pump oil can limit the ultimate
vacuum attainable. Pumps are available
with a gas ballast valve incorporated,
which minimizes the condensation of
vapors in the pump oil. The gas ballast
valve is a small valve that can be opened
manually to admit a controlled amount of
air to the pump cylinder during part of
the compression cycle. This will dilute the
vapors to the point where they do not
condense during compression. The violent
agitation of the oil by the additional air
rushing through the pump causes
reevaporation and exhaust of water that
may have been pumped from the vacuum
system in vapor form and condensed in
the pump oil. To effect the removal of
moisture when the surrounding air is
saturated with moisture, connect a dry
nitrogen gas supply to the gas ballast. Be
careful to select a nitrogen flow rate and
pressure that will not apply overpressure
to the casing of the pump. The extent of
use of the ballast valve is determined by
the amount of such vapors handled by
the pump. In normal high vacuum
service, the ballast valve is usually kept
closed because there is usually very little
water vapor present. The minimum
pressure obtainable is also slightly higher
with the ballast valve open. (Because of
the higher pressure in the chamber during
compression with the ballast valve open,
there is more leakage back into the
vacuum system.) Some pump
manufacturers recommend operating the
pumps with the ballast valve open once
each week for about 20 min to drive out
any water vapor that may have
accumulated in the pump oil.
Ultimate Pressure
Attainable in Rotary Pump
Vacuum Systems
The limiting absolute pressure approached
in a vacuum system, after sufficient
pumping time establishes that further
reductions in pressure will be negligible, is
called the ultimate pressure. The range of
ultimate pressures of commercial rotary
vacuum pumps extends from about 3 mPa
to 1 kPa (20 µtorr to 5 torr). The low
pressure of 3 mPa is reached only under
the most ideal conditions. The ultimate
pressure will be determined by:
1. outgassing of the pump,
2. the seal between rotor and stator,
3. contamination of pump oil and
4. the vapor pressure of the oil used.
A high vapor pressure pump oil will
evaporate at a greater rate, which will
create gas loads that saturate the pump
and limit the ultimate pressure attainable.
A disadvantage of any oil sealed and
lubricated pump is the backstreaming of
oil vapors from the pump inlet when inlet
pressures drop below or approach 70 Pa
(0.1 torr). This has become a major
concern to many industries, such as
semiconductor producers, for whom
backstreaming causes contamination of
their products with oil vapors. As a result,
several new pump designs classified as dry
or relatively free of this problem have
been available since the early 1980s. Two
of these are called scroll pumps and hook
and claw pumps.
Rotary Dry Mechanical
Pumps
Unlike the rotary vane pump, which
requires a low vapor pressure oil to
lubricate and seal the internal surfaces,
two commonly used pumps are designed
with very small clearances between the
moving and fixed surfaces and no need
for oil. As a result, the contamination
caused by vapors entering the evacuated
space at low pump inlet pressures is
Leak Testing of Vacuum Systems
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
225
eliminated. A slight disadvantage is that
these pumps cannot quite reach the same
low pressure as the lubricated pumps.
Ultimate pressures for dry pumps is in the
range of 2 to 3 mPa (20 to 30 µtorr).
Rotary Scroll Mechanical Pumps
A scroll is a free standing involute spiral
with a solid base on one side. A scroll set,
the fundamental element of a scroll
vacuum pump, is made up of two
identical right and left hand involutes.
When assembled, one scroll is indexed
180 degrees with respect to the other, to
allow the scrolls to mesh (Fig. 8).
In operation, one scroll is fixed and the
other is attached to an eccentric, driven
by an electric motor. The pump inlet is at
the periphery of the scrolls. As the
moving scroll orbits (but does not rotate)
about the fixed scroll, the entering gas is
trapped in two diametrically opposed,
crescent shaped pockets bounded by the
involutes and base plates of both scrolls.
The pockets shrink as they follow the
involute spiral toward the center,
compressing the gas. The compressed gas
exhausts to atmosphere through the
discharge port at the center of the fixed
scroll.
Rotary Claw Mechanical Pumps
The hook-and-claw mechanism consists of
several inline stages. The claw devices do
not make contact with each other or with
the chamber walls, obviating oils. Each
rotation of a claw pair consists of three
cycles: a start cycle, compression cycle
and a finish cycle. The two claws, which
divide the pump chamber, turn in
opposite directions and, in so doing, open
and close the intake and exhaust slots
through which the gases pass. During
Orbiting scroll
Pocket of gas
isolated
Gas inlets
shown
closed
Gas inlets
shown closed
226
Leak Testing
Pumping Speeds of Rotary
Mechanical Vacuum Pumps
Apart from the ultimate absolute pressure
that can be achieved by an particular
pump, there is an interest in how fast the
pump can reduce the pressure in a
vacuum system to the operating level.
Manufacturers normally specify the
pumping speeds of their mechanical
pumps at atmospheric pressure. In
general, rotary pumps start pumping at
atmospheric pressure and, as the pressure
is reduced, the pump becomes less
efficient. It then is pumping the same
volume, but at lower pressure. Eventually,
the pumping speed becomes zero at the
ultimate minimum pressure.
Figure 9 is a plot of pressure as a
function of pumping speed for a
400 L·min–1 (15 ft3·min–1) mechanical
pump. It is seen that at atmospheric
pressure, the pump is rated at
400 L·min–1; at 0.1 Pa (1 mtorr), the
pumping speed is 200 L·min–1 at 0.01 Pa
(0.1 mtorr), the pumping speed is
16.7 L·s–1 (35 ft3·min–1). The pump speed
reduces to zero at 10–3 Pa (10 µtorr), the
ultimate pressure attained by this pump.
At this point the gas handling capacity
has been saturated by the gas load from
the pump, thereby reducing its effective
pumping speed to zero.
Blower Pump or Booster Pump
FIGURE 8. Position of orbiting scroll shown before
compression cycle.
Exhausts at
center
opening
pumping, gas is drawn in one side of the
claws and compressed on the other as the
claws rotate. During rotation, the right
claw opens the intake slot, allowing gas to
be drawn into the chamber.
Simultaneously, the left claw opens the
exhaust slot letting compressed gas
escape. On completion of the
compression cycle, the claws pass through
a neutral position and cycle begins again.
The blower or booster pump (Fig. 10) is a
high throughput, low compression pump.
This pump is usually used on systems
where a large volume of gas must be
pumped. It is also used with a mechanical
pump to serve as the forepump for large
diffusion pumps, turbomolecular pumps
or even other blower pumps.
The pump consists of two figure eight
shaped rotors or lobes mounted axially on
parallel shafts, as shown in the drawing
below. These rotors are synchronized by
gears to prevent physical contact and
damage and rotate in opposite directions.
This rapidly displaces gas from the inlet to
the outlet.
Fixed scroll
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
How the Pump Works
Vacuum System Operation
These rotors are designed so that, while
spinning, they approach each other and
the housing within several thousandths of
an inch. (See Fig. 10.) Rotor speeds vary
from 40 to 60 s–1 (2500 to 3500 rotations
per minute). Because of the high speeds
and close tolerances of the rotating lobes,
booster pumps are usually not started
until roughing pressures of about 1.3 kPa
(10 torr) have been reached. The typical
blower windmills at atmospheric pressure,
producing much heat and very little
pumping action.
Blower or booster pumps are most
useful in the 0.1 to 0.01 Pa (1.0 to
0.1 mtorr) pressure range. They are always
backed by a mechanical pump as a result.
Operating at high pressures will cause
heating and expansion of the lobes. This
can result in damage to the pump. No oil
is used to seal the gap between stator and
rotor. Oil is used in the forevacuum
section of the pump to lubricate the gears
and bearings located there.
Operating procedure consists of turning
the mechanical pump on, then the blower
(Fig. 11). Usually the mechanical pump
has lowered the pressure sufficiently for
the blower to begin pumping by the time
the blower has reached operating speed. A
bypass valve around the blower is
sometimes used for high pressure
roughing.
Blowers are commonly used where
large volumes of gas need to be pumped.
They are used when the lowest pressure
needed is 10–2 to 10–3 Pa (75 to 7.5 µtorr).
They also are used to help the mechanical
forepump or backing pump maintain a
low pressure and help reduce the
possibility of oil backstreaming.
Turbomolecular Vacuum
Pumps
The turbomolecular pump serves as an
alternative to the diffusion pump and
must also be backed by a forepump. Its
Pump speed, L·s –1 (ft3·min–1)
FIGURE 9. Mechanical pump speed as a function of gas pressure for a pump rated at 6.7 L·s–1
(14 ft3·min–1) at atmospheric pressure.
7
(14.8)
6
(12.7)
5
(10.6)
4
(8.5)
3
(6.4)
2
(4.2)
1
(2.1)
Atmospheric pressure
10–3
10–2
10–1
100
101
102
103
104
105
(10–7)
(10–6)
(10–5)
(10–4)
(10–3)
(10–2)
(10–1)
(1)
(10)
Pressure, Pa (lbf·in.–2 × 1.45)
FIGURE 10. Blower pump operation: (a) at beginning of cycle; (b) after eighth of cycle; (c) after fourth of cycle; (d) after three
eighths of cycle.
(a)
Inlet
(b)
(c)
(d)
Outlet to
forepump
Leak Testing of Vacuum Systems
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
227
principle advantage over the diffusion
pump is that it provides an essentially
vapor free vacuum without baffles or cold
traps. Thus, for a system where the back
streaming of vapor from a diffusion pump
is undesirable or intolerable, a
turbomolecular pump could be used. Its
main disadvantage is that it has high
speed rotating parts whereas the diffusion
pump has not moving parts. It also
requires air gap tolerance on the order of
2 to 5 µm (8 × 10–5 to 2 × 10–4 in.)
between the high speed rotor and grooves
in the stator. As with a diffusion pump, a
molecular pump cannot operate at
pressures above 13 to 1.3 Pa (100 to
10 mtorr) and must be backed by a
mechanical forepump.
A turbomolecular pump (see Fig. 12) is
a mechanical vacuum pump that creates a
gas flow toward a suitable forepump by
imparting momentum or motion to gas
molecules by means of a rapidly rotating
rotor with successive rings with inclined
blades. These blades rotate with
circumferential speeds comparable to the
thermal motion of the molecules (speeds
of 100 to 700 m·s–1 or 330 to 2300 ft·s–1).
Some molecules are struck by the rotor
blades and rebound in a favorable axial
direction toward the stator blades. The
molecules rebound from these stator
blades in a direction favorable for their
being impelled by the next stage rotor
blades and so on as the process is repeated
through all successive stages of rotor and
stator blades. The series of impacts
statistically favor motion through the
turbine stages toward the discharge port
and constitute a pumping action with a
very high compression ratio.
The seal between the individual stages
is achieved by very narrow air gaps. The
dimensions of the grooves at the inlet
port must be such that the molecules
have a good chance of hitting the walls of
the groove or the blades without making
numerous collisions with other gas
molecules (see Fig. 12). As the gas is
compressed while passing through
successive stages of the turbine, it is
necessary to decrease the dimensions of
the air passages to keep them comparable
with the mean free path of the molecules.
The system must already be evacuated by
a forepump before a turbomolecular
pump can start pumping. It can achieve
pressures as low as 1.0 to 0.1 µPa (10 to
1.0 ntorr). Pumping speeds for air vary
from about 70 to 9000 L·s–1 (1.5 × 102 to
1.9 × 104 ft3·min–1), depending on the size
of turbomolecular pump selected.
Pumping speeds for hydrogen and for
helium vary only slightly from those for
air whereas the exhaust pressure is in the
range from 1.3 Pa to 1.3 mPa (10 mtorr to
10 µtorr). Higher exhaust pressures are
achieved in compound turbomolecular
FIGURE 12. Turbomolecular pump: (a) schematic; (b) inlet
port.
(a)
Gas inlet
Power source
for motor
To forepump
FIGURE 11. Vacuum system with blower
pump.
Chamber
Roughing valve
(b)
High
vacuum
valve
Blower pump
Stator
Mechanical pump
Rotor
228
Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
pumps. These follow the turbomolecular
stages with one or several molecular drag
stages, which further compress the gas
through the effects of viscosity.
Operation of Vapor or
Diffusion Pumps for
Vacuum Systems
Although mechanical rotary pumps
sometimes attain pressure below 0.1 Pa
(10–3 torr), they are generally used in the
100 to 0.1 Pa range. To obtain pressures
well below 0.1 Pa, the vapor pump was at
one time the most commonly used.
However, in the 1990s it was largely
replaced by turbomolecular pumps
because of the backstreaming of vapors.
The principle of operation of vapor
pumps is entirely different from that of a
rotary oil sealed pump, where the gases
and vapors are compressed by a rotating
mechanical member and exhausted to the
atmosphere.
The vapor pump, or diffusion pump,
operates in the molecular flow region. The
basic principle involved is shown in
Fig. 13.
The pump works by heating the pump
fluid to its boiling point. The vapors
travel upward inside the jet assembly and
exit through the jet nozzles. In fact, they
are accelerated downward through the jet
nozzles. The vapor molecules travel very
fast and can reach supersonic speeds.
FIGURE 13. Principle of operation of high vacuum vapor
pump. Vapor forced through a narrow opening (nozzle)
attains a high speed and is directed at a downward angle.
Molecules of gas or vapor that wander along a path toward
the jet stream will be struck by vapor molecules. The gas
molecule B has diffused into the path of the jet stream where
it is struck by the vapor molecule A. Molecule B is given a
generally downward motion.
These vapor streams are directed
toward the outer walls of the pump. The
walls are typically cooled by water. When
the vapor hits the cooled walls, it
condenses back into a fluid. This fluid
then flows downward into the pump
boiler for reboiling.
The actual pumping of gases happens
when the large, heavy, high speed oil
vapor molecules hit gas molecules. The
gas molecules are knocked downward and
compressed by the movement of the
vapor jet stream.
The gas molecules are thereby
compressed in several stages to higher
pressures. They are finally pumped away
through the foreline by the mechanical
pump (Fig. 14). When the oil drops to the
bottom of the pump, it is reboiled and the
cycle repeats.
Vacuum Limitations of
Vapor Diffusion Pumps
A diffusion pump (Fig. 14) cannot
operate at pressures above 0.1 Pa (1 mtorr)
because the oil vapor jets cannot form in
the viscous flow region. Therefore, the
pump must start pumping in a chamber
that is already under vacuum (such as
that attained with a rotary mechanical
forepump). Oil is the most frequently
used diffusion pump fluid because of its
low vapor pressure at room temperature.
Oil has a fairly steep curve relating its
pressure to temperature. This is necessary
for proper operation of the pump boiler.
The lowest attainable pressure of the
diffusion pump is determined in part by
FIGURE 14. Construction of three-stage high vacuum vapor
pump.
Cold cap
Inlet
Multistage
jet assembly
Flow from
vacuum chamber
Cylindrical
water cooled
body
Exhaust
Nozzle
Thermal
protect
switch
Baffles
Foreline
A
Ejector
B
Electrical connector
Fill and
drain
assembly
Vapor
Vapor jet
Oil reservoir (boiler)
Heater
Leak Testing of Vacuum Systems
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
229
the vapor pressure of the oil at the
temperature of the available cooling
water. Oils specified by pump
manufacturers have vapor pressures,
under these conditions, of about 0.1 µPa
(1 ntorr). The popularity of the diffusion
pump is due to its wide range of
operation, low cost, reliability and lack of
moving parts. The pump heaters are
usually mounted from the outside and
can be replaced during operation.
Should the diffusion pump be suddenly
exposed to a burst of atmospheric
pressure, the oil jet stream would collapse,
thereby destroying the pumping
capability of the vapor pumps and
possibly acting to crack the oil. The term
cracked oil refers to a decomposition of
pump oil due to exposure to oxygen in
the atmosphere while at or near the
boiling point of the oil. Some fluids are
less susceptible to cracking than are other
diffusion pump fluids.
Operation of Baffles and
Traps in Vacuum Pumping
Systems
One of the objections to diffusion pumps
has been the possibility of contaminating
the vacuum chamber work area with the
pump fluid. By providing suitable traps
and baffles between the pump and the
vacuum chamber, back diffusion of oil
and oil vapor can be minimized and
condensable vapors from the chamber
may be trapped. As a general rule, the
pumping speed of the system goes down
as the trapping efficiency of baffles and
traps goes up, due to decreased
conductance.
The baffle or trap should normally be
kept as cold as possible. However, the
temperature of surfaces of the first
baffling state above a pump should be
cool enough to condense the oil vapors,
but not so cold as to freeze the pump oil
and prevent it from flowing back into the
pump.
of gas, it should not sacrifice high
conductance because that would impair
the net pumping speed of the system.
Operation of Cold Traps in
Vacuum Pumping Systems
A cold trap placed above the baffle
ensures that those few oil molecules that
may get by the baffle will not get to the
vacuum chamber. A cold trap, therefore,
stops back migration of pump oil vapors.
It is also very effective as a cryogenic
pump for pumping condensable vapors
such as water vapor, the chief offender in
most systems, as well as for grease vapors
and other undesired contaminants.
As a cryogenic pump, the cold trap
reduces system pressure by taking
molecules out of the gas or vapor phase
and trapping them on its surface. These
molecules are not pumped out of the
vacuum system and discharged to
atmosphere. The most common
techniques used to obtain low
temperatures for cold traps are mechanical
refrigeration, dry ice and liquid nitrogen.
Some common forms of optically dense
chevron and cold traps are shown in
Fig. 16, which also shows thimble type
traps used in mass spectrometer leak
detectors. The reservoir is filled with
liquid nitrogen through the filler tube.
Use of liquid nitrogen requires that the
thimble type trap be kept essentially in a
vertical position.
Characteristics Desired in Vacuum
Valves
Vacuum valves must (1) be free from
leakage, (2) offer minimum flow
FIGURE 15. Typical baffle designs used in oil diffusion vacuum
pump systems.
Plate
Cooling coils
Cooling
coils
Characteristics Desired in
Diffusion Pump Baffles
A baffle is simply a cool surface that is
placed above the diffusion pump in the
path of gas flow. This baffle is of metal
with good thermal conductivity that
keeps its surface at a uniform temperature.
The refrigerant, usually cold water, is
passed through tubing that is brazed to
the baffle. A baffle should also be
optically dense, that is, there should be
no line of sight through it, to avoid back
flow of molecules in molecular flow.
Fig. 15 shows some typical designs of
baffles. Because a baffle restricts the flow
230
Leak Testing
Top view
of disk
Cooling
coils
Cooling coils
Cooling coils
Top view
of chevrons
Cooled chevron trap
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
resistance and (3) contain materials that
do not outgas. The biggest problem in
making leaktight valve is in sealing the
operating shaft. Two types of valves that
accomplish efficient sealing are the
bellows sealed and diaphragm valves
(Fig. 17). Usually preferred are brass or
stainless steel bellows, more movement
being obtained with brass. The bellows is
brazed to the cover (bonnet) and dish, as
shown in Fig. 17a. Figure 17b shows a
valve using a diaphragm that can be a
metal or elastomer. Compared to metal
diaphragms, an elastomer has
considerable flexibility but also has the
disadvantages of outgassing and
permeability to various gases. On the
other hand, metal diaphragms are not as
elastic but have better outgassing and
permeability characteristics.
Precautions in Disassembly of
Bellows Sealed Valves
Always open bellows valve before
removing the stem assembly to prevent
cracking the bellows. Never completely
extend bellows when out of the valve.
Operation of Capture
Vacuum Pump
Unlike the previously described pumps,
which compress and exhaust gas either to
atmosphere or into an attached forepump,
two commonly used pumps collect and
store gases in the pump body until
eventually being released to atmosphere
by a process called regeneration (for the
cryopump) or until the pump is rebuilt as
in the ion pump. These pumps are the
mechanical cryopump and the sputter ion
pump.
FIGURE 17. Operating principles of vacuum valves:
(a) bellows sealed valve; (b) diaphragm valve.
(a)
Cover
FIGURE 16. Cold traps used in vacuum pumping systems to
condense vapor molecules: (a) combination baffle and trap
with optically dense chevrons; (b) thimble trap used in leak
detectors.
Bonnet gasket
Braze
(a)
Bellows
Liquid
nitrogen
Body
Water
Seat
(b)
Liquid nitrogen
(b)
Braze
Vent hole
Diaphragm
To
chamber
Braze and mechanical seal
if an elastomer diaphragm
To diffusion
pump
Leak Testing of Vacuum Systems
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
231
Operation of Cryopump
The cryopump is unique in that it pumps
by getting the gases so cold that they
freeze and are stored, or captured, in the
pump. It is extremely clean, using no oil
and having no moving parts in vacuum.
It also has a very high throughput and is
used in the high vacuum range in
industrial applications where
hydrocarbons cannot be tolerated.
A cryopump (Fig. 18)2 is made up of
two main components: a gaseous helium
compressor and a pump consisting of an
expander, cold head (chilled surfaces) and
the pump body. These two components
are connected by flexible hoses to form a
closed loop refrigeration system. Gaseous
helium is circulated between the
compressor and expander.
The pump module consists of the
expander module, the first and second
stage cryoarrays, the pump body, second
stage temperature monitors and a pressure
FIGURE 18. Schematic of cryopump.
C
D
E
F
G
H
I
A
J
B
Legend
A = Forevacuum port
B = Power connection
C = Inlet flange
D = Baffle
E = Second cold stage
F = Radiation shield
G = Cryoplates
H = Relief valve
I = First cold stage
J = Vapor pressure thermometer
232
Leak Testing
relief valve. In the expander, high pressure
helium is supplied by the compressor.
This gas is expanded in two stages to
produce cryogenic temperatures. The
actual operating temperatures will vary,
depending on the thermal and gas loads
that are imposed. The first stage operates
between 50 and 80 K (–370 and –315 °F)
and the second stage between 10 and
20 K (–440 and –420 °F). The cryoarrays
are the pumping surfaces, cooled by the
expander, on which gases from the
vacuum chamber are condensed or
adsorbed.
In cryopump operation, helium is
compressed and gives up its heat to the
surrounding walls of the compressor. This
heat is removed by water or air cooling.
The cooled, compressed helium then goes
to the pump cold head. The expander at
the cold head valving system lets the
helium expand. The expanded helium
now absorbs heat from the cold head and
baffle array. This chills the cold head and
baffle array to about 12 K (–440 °F) and
70 K (–335 °F), respectively. These chilled
surfaces pump gases from the vacuum
chamber in two ways. The gases are either
condensed or adsorbed on the arrays.
That most gases will stick to a surface in
an icelike state at less than 20 K (–420 °F)
is very likely. At this temperature, the
combination of partial pressures of most
gases is about 10–9 Pa (10–11 torr) or lower.
Most gases are condensed on the first
and second stage cryoarrays. The first
stage array is cold enough to pump water
vapor and carbon dioxide by
cryocondensation. The colder second
stage array pumps nitrogen, oxygen,
argon and most other gases by
cryocondensation, but is not cold enough
to condense helium, hydrogen and neon.
These three gases are pumped by the
process called cryosorption; a surface
related phenomenon: the greater the
available surface area at cryogenic
temperatures, the more likely that gas
molecules will stick to it. Although most
gases are frozen or condensed between
12 and 20 K (21 and 36 °R), helium,
hydrogen and neon are still very actively
in motion at these temperatures. If we did
not remove them, their partial pressures
would continue to rise, perhaps to a point
where the total system pressure would be
unacceptable.
To solve this problem, activated
charcoal is attached to the bottom side of
the second stage (coldest) cryoarray where
it is less likely to adsorb the easier to
pump condensible gases. This reserves the
charcoal for the helium, hydrogen and
neon which are trapped in the maze like
structures and surfaces of the charcoal.
This is similar to a sponge soaking up
water vapor at room temperature. This
process is called cryosorption.
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
Before chilling the cryoarrays, the
pump volume must be rough pumped to
remove most of the gas below a valve at
the pump inlet. During chilling, when the
second stage drops to less than 20 °K, the
pump is ready for use. During use, the
pump can absorb very large amounts of
condensible gas, but the second stage
charcoal eventually saturates, usually with
hydrogen and must then be turned off to
warm up the pump (regeneration). To
speed up this process, dry nitrogen is
applied to the purge tube, through a
valve, which flushes the pump and expels
the previously stored gas out through the
pressure relief valve. When the second
stage rises to room temperature, the pump
is ready to be rough pumped and chilled
again.
The cryopump is normally used in the
pressure range of 10–1 to 10–6 Pa (10–3 to
10–8 torr) but when operated continually
at the upper end of this range, the time
between required regeneration cycles is
proportionately shorter; i.e. — more
downtime.
are accelerated toward the anode. This
long path increases the probability of
ionization and therefore the amount of
useful pumping action that can be
performed by the pump.
Because of the action of the magnetic
field, the electrons do not easily come in
contact with the anode. As a result, a
cloud of electrons is formed within the
anode space. This electron cloud becomes
fairly stable during pump operation and is
dense enough for the efficient ionization
of gas molecules. The name for this
process is cold cathode discharge. The
positively charged ions, which are
relatively heavy particles, are accelerated
into the negatively charged titanium
cathodes. This impact causes sputtering,
or chipping away of the titanium cathode
material.
Sputtered titanium deposits onto the
internal structure of the anode. Then,
when gas molecules come in contact with
these clean titanium deposits, chemical
Operation of Ion Pump
FIGURE 19. Section through a cold cathode ionization gage
(Penning gage).
The ion pump (Fig. 19)2 is also a gas
capture pump but is not designed to
pump heavy gas loads. For this reason, it
is not generally used alone in high
production applications. It is more often
used in research and analytical
applications where there is no need to
cycle the work chamber repeatedly and
rapidly from atmosphere to vacuum.
Ion pumps are clean operating
electronic devices which use no moving
parts or oils within the vacuum pump. It
is possible to achieve pressures in 10–9 Pa
(10–11 torr) range with overnight bakeout
of the system. The bakeout process drives
residual gas off the system walls, which is
then pumped by the ion pump. In
research and analytical applications, the
ion pump’s cleanliness, bakeability, low
power consumption, and long life make it
the pump of choice for most ultrahigh
vacuum uses. They are available in various
sizes and variations, but only the simplest
(diode) pump will be described here for
purposes of brevity.
A stainless steel ion pump body
contains a multicell anode assembly
constructed of cylindrical parallel tubes
spaced between two flat titanium
cathodes. A very strong magnet is placed
outside the pump body. After the ion
pump is rough pumped to 1 Pa (10–2 torr)
or less, a voltage of 5 to 7 kV direct
current is applied between the cathodes
and the anode assembly.
The magnetic field forces any free
electrons within the anode into long
helical paths instead of straight paths.
This increases the probability of electron
collision with molecules, as the electrons
A
H
B
I
J
C
K
D
L
E
M
F
N
G
Legend
A = High voltage connection
B = Hood
C = Protective cap
D = Vacuum tight cast iron housing
E = Permanent magnet
F = Small flange connection
G = Baffle
H = Safety terminal
I = Leadthrough (anode lead)
J = Compressed glass-to-metal seal
K = Ring anode
L = Ignition pin
M = Fixing screw
N = Cathode plate (exchangeable)
Leak Testing of Vacuum Systems
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
233
combination converts these gas molecules
to solid compounds such as titanium
oxide or titanium nitride. This process is
called chemical gettering and produces
the required pumping action. In addition,
a second pumping action takes place.
Some of the ionized molecules impact the
cathodes with enough force to become
buried in them, which prevents them
from neutralizing and becoming a free gas
again. A third pumping action occurs
with hydrogen which diffuses directly
into and reacts with the cathode plate.
Also, neutral particles such as the inert
gases can literally be buried or covered by
the sputtered cathode material. Complex
molecules may also be split apart in the
discharge to smaller, more readily pumped
molecules. Because these actions are not
equally efficient, the chemically reactive
gases such as hydrogen, nitrogen and
oxygen are pumped at much higher
speeds than the inert gases. A
modification of the cathode design can be
made to increase the efficiency for these
inert gases.
Another characteristic of the ion pump,
often referred to as a sputter ion pump, is
that it is self-regulating. At higher
pressures, where much ionization takes
place, more current flows and at low
pressures, less current flows. This
characteristic current drain can be used to
measure the pressure, or degree of vacuum
achieved with the pump. This feature
eliminates the need for an ion gage on
FIGURE 20. Schematic diagram of a typical complete vacuum
system.
Pumping
port
Bell jar chamber
or
other process vessel
Vent valve
High
vacuum
valve 1
Roughing
valve 2
Cold trap
Baffle
Foreline
valve 3
To
atmosphere
Diffusion
pump
Ballast
tank
Leak test
port
Mechanical pump
234
Leak Testing
the system. At lower pressures, ion pumps
have long lives. Once they begin
pumping, they quickly lower the pressure
to the long life region. As long as they are
not pumping against a leak, they will last
for years. An example of this would be
that a pump working at a constant
pressure of 10–5 Pa (10–7 torr) would have
a useful life of 20 years.
Procedures for Pumping
and Operating Complete
Vacuum Systems
By combining the components previously
discussed with appropriate manifolding,
plumbing and gaskets (O-rings), a
complete vacuum system may be built as
shown schematically in Fig. 20. The initial
conditions are:
1. mechanical pump running,
2. diffusion pump operating and working
in high vacuum,
3. cold trap filled with liquid nitrogen,
4. atmospheric pressure in bell jar
chamber,
5. high vacuum valve closed,
6. vent valve open,
7. roughing valve closed and
8. foreline valve open.
An operational cycle for this vacuum
system is as follows:
1. Close access to bell jar chamber, vessel
or hood to be evacuated.
2. Close vent and foreline valves. The
ballast tank permits the
turbomolecular pump or diffusion
pump to discharge to an expansion
volume so that a high critical
forepressure is not reached.
3. Start the roughing cycle by opening
the roughing valve. This allows the
mechanical pump to evacuate the
manifolding between the high vacuum
valve and the bell jar chamber.
4. After the pressure has been reduced to
below 10 mPa (about 50 µtorr) or
crossover point, close the roughing
valve.
5. Open the foreline and high vacuum
valves. This allows the diffusion
pumping system (cold trap, baffle and
turbomolecular diffusion pump) to
continue pumping until the desired
operating pressure is reached and work
in the chamber may commence.
After completion of work in the bell jar
or vacuum chamber, the system may be
cycled to its initial condition by first
closing the high vacuum valve and then
opening the vent valve. This allows
atmosphere to enter the bell jar chamber
and system up to the roughing and high
vacuum valves. The pressure equalization
allows access to the chamber.
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
PART 3. Materials for Vacuum Systems
Outgassing of Materials in
Vacuum Systems
Adsorption refers to the condensation of
gas (vapor) on the surface of a solid. As
the pressure is reduced in a vacuum
chamber, there is a spontaneous evolution
of gas (and vapor) from materials in the
vacuum; this is referred to as outgassing.
In vacuum systems the materials in the
vacuum region may release adsorbed gases
and vapors that increase the gas load of
the system, resulting in a much longer
pumpdown time. This phenomenon is
most prevalent in new vacuum systems,
unclean vacuum systems or vacuum
systems that have been exposed to
atmosphere for some considerable time. It
will also occur when new materials or
new work jigs and fixtures are installed in
a vacuum chamber. Knowledge of the gas
adsorption properties of various materials
and, therefore, their outgassing properties,
is very valuable in vacuum work.
Technique for Releasing
Adsorbed Gases by
Moderate Heating
Most metals in vacuum give off
adsorbed or dissolved gases as well as
gases resulting from the decomposition of
oxide near the surface. To minimize this
gas evolution, metals can be heated under
vacuum before being used in vacuum
systems. Gas adsorbed by exposure to
atmospheric pressure can easily be released
by heating to moderate temperatures.
When pumping to pressures below 0.1 mPa
(1 µtorr) where baking is not practical, great
care must be taken in choosing the various
materials in the system. This applies to
choice of vacuum greases, elastomers,
metals and various sealing compounds.
Factors Influencing
Adsorption and
Outgassing by Baking
Vacuum System Materials
Gases and vapors are adsorbed by vacuum
construction materials (metals and
elastomers) and are gradually released.
This set one limit on the lowest ultimate
pressure that can be reached in a
particular vacuum system. The usual
technique of overcoming this problem is
to degas the materials, usually by baking
(raising the system to a high temperature
while pumping). The bake-out
temperature will depend on the
temperature at which the material begins
to change its properties. consequently,
vacuum systems are degassed at fairly
modest temperatures, say 300 to 400 °C
(570 and 750 °F), for several hours while
being pumped. This will eliminate much
of the adsorbed gases and vapors.
The dissolved gas content of a metal or
alloy will depend on factors such as
(1) the nature of the metal, (2) the
metallurgical process used in the
production of the metal and (3) the
degreasing and cleaning to which a metal
was subjected. In comparing metals such
as stainless steel and aluminum, stainless
steel is found to outgas at a much lower
rate. A cast aluminum surface outgasses at
a rate about ten times higher than the
rate at which a stainless steel surface
outgasses. Therefore, during vacuum
pumping, stainless steel vacuum systems
are capable of reaching a desired vacuum
in a shorter time than a comparable
aluminum system with its higher rate of
outgassing. Results are strongly influenced
by the condition of a metal (its alloy,
cleanliness, finish etc.).
Functions of Elastomers as
Gaskets and Seals in
Vacuum Work
Certain openings must be provided for the
insertion, removal and sealing of
equipment or materials for a given
vacuum system. During the operation,
these openings must be tightly sealed.
Elastomers are the most widely used gasket
material, where temperature and gas loads
permit, because they offer reliable sealing.
Elastomers are natural or synthetic rubbers
that can be vulcanized to a state in which
they have an inherent ability to accept
and recover from extreme deformation.
For high vacuum service, leaks must be
entirely eliminated and the gas evolved
from the gasket material itself must be
negligible. Both natural and synthetic
rubber satisfy these requirements as long
Leak Testing of Vacuum Systems
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
235
LT.06 LAYOUT 11/8/04 2:17 PM Page 236
as negligible surface area is exposed.
Frequently, gaskets must be exposed to oil
or other gasket deteriorating substances
and sometimes rather high or low
temperatures must be tolerated. Gasketed
joints should be readily accessible for tests
for leakage. In designing a gasket these
factors must be considered and
specifications should be based on material
capabilities as well as vacuum system
operational requirements.
Selection of Gasket
Materials and Design for
Vacuum Seals
The choice of natural or synthetic rubber
for a vacuum application depends on the
combined qualities desired. In the case of
rubbers, a wide range of characteristics is
acceptable. Perhaps the most important
single factor is that of allowable deflection
under compression. This is a function of
hardness and allowable permanent set.
These materials generally contain volatile
oils, plasticizers and coloring pigments
that adsorb moisture and gases. Most of
the chemicals used have low vapor
pressure at room temperature. The
outgassing rates for various elastomers
depend on factors such as (1) the
formulation used, (2) the area exposed,
(3) the operating temperature and (4) the
treatment of the elastomer before use. As
a rule, there is no way to control the
formulation of gasket materials because
this is determined by the manufacturer.
However, it is feasible to inform the
manufacturer of intended service and ask
for minimum volatiles. Exposed gasket
area becomes critical as the operating
pressure is lowered. Proper gasket groove
design can help considerably in reducing
exposed areas.
Because the outgassing rate of
elastomers increases as the temperature is
raised, the ultimate pressure can be
reached more rapidly if the elastomer can
be heated. However, all elastomers are
damaged when heated too much. Also,
the compression set increases more
rapidly with temperature.
Because of these properties, elastomeric
gaskets are not normally used in ultrahigh
vacuum systems. Such systems are baked
at temperatures well above the damage
point of all known elastomers. In this
case, it becomes necessary to use joints
and seals of metals and alloys such as
aluminum, brass, bronze, copper, indium,
lead, silver, stainless steel and others.
Properties of Specific Elastomers
for Vacuum Seals
Natural and synthetic rubbers are
commonly used in systems that operate at
236
Leak Testing
room temperature and at pressures near
1 mPa (10 to 1 µtorr). Because of its
temperature tolerance, silicone rubber is
commonly used for low and high
temperature operation. The fluorinated
elastomers are highly resistant to most
corrosive materials found in vacuum
practice. Fluorocarbon resin is very good
but suffers from cold flow under pressure
at room temperature; suitable means for
containing the fluorocarbon resin (spring
loaded gaskets etc.) will eliminate this
difficulty. In trying to reach very low
pressures, the permeability of the elastomer
as well as its outgassing characteristics must
be considered. Permeability is the property
that determines how readily gases will pass
through a material.
Selecting Elastomers to Reach Low
Pressure Vacuums
To reach low pressures at room
temperature, elastomers with low vapor
pressures and low permeabilities are
desirable. consequently, considerable work
has been done with fluorinated elastomers.
Baking an elastomer at a temperature that
does not damage it will reduce pumpdown
time; however, it will still release some
vapor after many hours of pumping.
Selection of Alloys for Use
in Vacuum System
Components
There are many alloys of copper, but only
brasses and bronzes are used in vacuum
practice. Brasses are copper zinc alloys,
whereas bronzes are copper tin alloys.
However, many brasses contain various
other metals. Brasses are widely used for
vacuum parts, such as diffusion pump
parts, chambers, base plates, valves and
fittings in high speed dynamic vacuum
systems.
Many commercial bronzes contain
zinc. Alloys containing zinc, cadmium,
lead, antimony or bismuth should not be
used in vacuum systems that are to be
baked because of the high vapor pressures
of these metals. Vacuum firing is likely to
alter the composition and therefore the
properties of such alloys.
Properties of Austenitic Stainless
Steels in Vacuum Systems
Stainless steels have come into fairly
common use in vacuum practice for
turbomolecular pumps, diffusion pumps,
manifolds, chamber baseplates etc.
Austenitic stainless steels (types 302, 303
and 304) are commonly used in vacuum
work and are often called 18-8 stainless
steels because they contain about
18 percent chromium and 8 percent
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
nickel. These steels are nonmagnetic and
the melting points of austenitic stainless
steels are over 1400 °C (2550 °F). Surfaces
of stainless steels remain smooth because
oxides and hydroxides do not occur as in
other types of metals. This means that the
effective surface area is less and vapors are
adsorbed in smaller quantities. This leads
to much easier degassing and quicker
pumpdown.
Properties of Aluminum Alloys
Used for Vacuum System
Components
Aluminum is also being used in vacuum
systems. The alloys of aluminum are
generally readily worked in the shop
without much difficulty, the workability
depending on the composition. Surface
hardening can be achieved easily by
anodizing and other processes. Parts may
be joined together by using aluminum
solder. Cast aluminum alloy parts are used
for a variety of purposes such as valves,
turbomolecular pumps, diffusion pumps
(particularly jet assemblies), grooveless
flanges and gaskets. The design of the dies
is important to get vacuum tight
aluminum die castings. Although
aluminum is difficult to de-gas
thoroughly, it is commonly used for
vacuum parts where good heat and
electrical conductivity is required.
Properties of Other Metal Seals in
Vacuum Systems
Certain specialty metals have almost the
same coefficient of expansion as most
glasses and have excellent sealing
characteristics. They are used with
vacuum flanges in the manufacture of
ionization gage tubes and in other
applications where metal-to-glass
junctures and seals are necessary.
Applications and
Limitations of Soft Metallic
Vacuum Gaskets
Metal gaskets of some kind are used by
vacuum seals that must be maintained at
temperatures higher than about 125 °C
(257 °F) or in which rubber cannot be
used because of outgassing. Small gaskets
of lead, copper, aluminum, gold, silver or
tin have long been used for higher
temperature vacuum services. Complete
sealing demands high stresses and
consequently the metal gaskets can only
be used once. They are not designed for
applications where the seals are often
opened and then reclosed because the
metal gaskets will take a permanent set
and are not reusable in most applications.
Selection and Properties of
Vacuum Greases and Oils
Vacuum greases are commonly used to
help attain seals and to lubricate devices
such as stopcocks and gasketed joints
(static, rotating and sliding). In some
cases, vacuum oils are used, including
diffusion pump oils. Oils are generally not
as satisfactory as greases for most types of
seals, because they are more readily
squeezed out, thereby leaving a dry seal.
In general, vacuum greases should not
have a vapor pressure of more than about
10 mPa (0.1 mtorr) at 30 °C (86 °F) and
should maintain adequate viscosity at this
temperature and can be used up to a few
degrees below their melting point. In
general, vacuum greases should be applied
sparingly and surplus grease then wiped
off, because greases absorb gases and
vapors and are dirt catchers.
Diffusion Pump Oils
The ultimate vacuum of many vacuum
systems is, in fact, limited by insufficient
trapping of gas molecules by the diffusion
pump fluid. Certain desirable properties
that a diffusion pump oil must have
include the following.
1. It should have low vapor pressure.
Vapor pressures of typical diffusion
pump oil recommended by
manufacturers of diffusion pumps are
in the range from 10 to 0.01 µPa (100
to 0.1 ntorr).
2. It should have low enough viscosity to
flow back into the boiler.
3. It should have high molecular weight
relative to the pumped gases to
increase the efficiency of removal of
gas from systems being evacuated by
the vapor jets. Molecular weight of oil
commonly used is in the range of 300
to 500 unified atomic mass units (u).
4. Oil should be thermally stable to
avoid decomposition with heat.
Decomposition often results in the
evolution of more volatile fractions
caused by cracking of the oil due to
frequent exposure to atmospheric
pressures.
5. The fluid should be chemically stable
and noncorrosive in the presence of
common metals, glass, elastomer
gaskets and the gases and vapors
usually present in vacuum systems.
6. It should be nontoxic.
The recommended hydrocarbon oils
represent a very satisfactory low cost fluid
for the normal vacuum range down to the
low 10 µPa (100 ntorr) region. Ultrahigh
vacuum is best obtained with oils specified
by pump manufacturers. These oils are
extremely stable, showing little change in
properties even if the pump is exposed to
atmospheric pressure with the heater on.
Leak Testing of Vacuum Systems
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
237
PART 4. Vacuum System Maintenance and
Troubleshooting
Maintenance of Vacuum
Systems
The recognition, diagnosis, troubleshooting
and treatment of vacuum system
malfunctions and analysis of specific
problems such as leaks commonly
encountered in any vacuum system are
important factors in maintaining vacuum
systems at satisfactory levels of
performance. The amount of maintenance
service required by a vacuum system will
depend on three basic factors:
1. The cleanliness of objects to be vacuum
processed. Objects that are to undergo
evacuation should be thoroughly
degreased. Compounds or lubricants at
connection points within equipment
should always be held to a minimum.
2. The physical environment of the
entire vacuum system. A clean
temperature controlled environment is
highly conducive to a long trouble
free life of any vacuum system.
Extreme ambient temperatures or high
residual dust levels can appreciably
affect the degree of trouble free
operation to be expected from the
system. When setting up a preventive
maintenance schedule for any vacuum
system, the actual environment in
which the system is expected to
function should be given prime
consideration when selecting the rates
and/or scheduled times at which
specific preventive maintenance is
performed. Under the heading of
physical environment, one should also
consider very carefully the reliability
of available air, water and power
sources. Although many vacuum
systems are protected adequately
against most emergencies, air, water or
power failures with any vacuum
equipment do not contribute to the
overall well being of the machine.
3. The human element. The most serious
consideration in maintenance of
vacuum systems is that of personnel
experience, care and training. Even
with self-protected automatic vacuum
machines, breakdowns do occur. If a
unit is of the manual variety,
particular concern should be directed
to the human element. One cannot
take too many precautions to prevent
unauthorized personnel from
tampering with a high vacuum
238
Leak Testing
evaporator or pumping station. This
should be recognized in
troubleshooting because it may well be
the cause of certain problems.
Selecting Vacuum System
Operating Schedules to
Reduce Maintenance
Maintaining the cleanliness of internal
machine parts exposed to high vacuum
requires that the pumping system of a
unit be kept running continuously as a
machine cleaning function. In addition,
the liquid nitrogen cold trap should not
be permitted to run empty over night and
over weekend periods. On manual as well
as semiautomatic systems, strict attention
should be paid to the proper
manipulation of the system valves and to
the selection of personnel having access
to these valves. If the entire system has
undergone cleaning, it is advisable to
permit it to operate for a 24 h period
without liquid nitrogen in the cold trap
and with the port to the chamber or test
volume blanked off. The preceding
comment applies, although to a lesser
degree, whenever the actual high vacuum
portion of the system, i.e., that part of the
system beneath the high vacuum valve,
has seen atmospheric pressure, whether
intentionally or otherwise, for more than
a very brief period of time.
Delegating Responsibility
for Operating Vacuum
Systems
The human element problem is
something best worked out within the
individual company or group responsible
for the vacuum system. Generally, it
would seem best to delegate total
responsibility for the operation and
maintenance of the vacuum system unit
to one responsible individual. Field
experience tends to indicate that far fewer
field problems occur with equipment that
is owned and maintained under well
defined levels of responsibility. Far more
servicing is required for vacuum systems
where no specific individual or group is
held directly accountable for the
condition of the equipment. Automation
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
of startup and operating sequences
minimizes these problems. Contractual
service agreements can usually be
obtained for the routing servicing and
maintenance of vacuum equipment.
Preliminary Techniques for
Locating Faulty conditions
in Vacuum Systems
Frequently, maintenance checks show that
the existing trouble with vacuum systems,
although real enough, is not actually the
result of a machine part failure.
Consequently, on the assumption that the
equipment was operating satisfactorily up
to the point of failure, the following
procedures for checks of basic power, water
and air supplies should be followed:
1. Using a volt meter, check to make
certain that the specified voltage is
available at the power electrical outlet
being used. Frequently, circuit breakers
are opened within a plant.
Occasionally workmen make power
wiring changes within a plant and
inadvertently disable parts of the
electrical system. The operator should
not assume that power is available at
the wall receptacle unless he or she
has personally checked and proven
that the power is present.
2. If necessary, disconnect the outgoing
water line from the system and be
absolutely sure that cooling water is
flowing through the water cooled
component and exiting to the drain.
Occasionally the water circuit will
become plugged by debris in the line.
Because some machines are protected
against temperature rise in the
diffusion pump, only roughing level
vacuum may be achieved due to the
automatic turnoff of the diffusion
pump because of improper water flow.
If the water flow is found to be
blocked, correct this condition and
continue with the machine startup
procedure as specified in the
manufacturer’s operating instructions.
3. After checking water and power, be
sure that proper air pressure is being
maintained for actuating air operated
valves. Low air pressure can cause
some rather strange operational
symptoms, which may be
misdiagnosed as a vacuum controller
failure or sticky valves. As often as not,
low air pressure is the cause of
sluggish or nonfunctioning valves.
4. Startup procedures should be reviewed
to make certain that all operational
switches are properly set and that the
unit should indeed be running
normally. No matter what the visible
trouble symptoms may be, the
aforementioned procedures should be
followed before other service
procedures are attempted.
Because power, water and other utilities
vary considerably with the type of pumps
and systems being used, the previous
suggestions are only general. For more
specific information, refer to the
manufacturers instruction manuals.
Providing Necessary
Information to Service
Engineers
If, after completing the basic air, water
and power checks described above, a
simple explanation for the machine
malfunction is not found, a written record
should be prepared covering the following
information:
1. A statement covering the age and
history of the vacuum system, the
serial number, what it has been used
for, what it is currently being used for,
who used it and in what manner,
types of materials being used in the
vacuum system, available maintenance
history and in general, as many details
as can be acquired.
2. Note carefully the symptoms observed
with the particular machine and what
has been done to this point about
correcting these problems. When this
information is available, do not
hesitate to call the service engineer for
the equipment and give him all details
possible. It is entirely possible that,
given useful information, he or she
may be able to prescribe, via phone,
the course of action needed to cure
the vacuum system’s troubles.
Also, if thorough information can be
acquired via phone, the service engineer
will be much better prepared to take care
of the problem when he or she arrives at
the plant, should that be necessary. The
time it takes to repair the system will
often be a function of the quality of
communication between the plant and
the service engineer.
Selecting Service Personnel within
User Organization
Whether a service engineer has been
called or not, if it is preferred to proceed
immediately with troubleshooting a
vacuum system, it may be possible to
arrange for the services of a qualified
individual within the user organization.
Generally, the first choice for
troubleshooting should be someone
within the company who has had
previous vacuum system experience
whether with the same type of equipment
Leak Testing of Vacuum Systems
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
239
or with some other type. A large
organization may have a complete
department devoted entirely to the
maintenance of vacuum equipment. It is
also possible that within a company some
individual may have responsibility for
maintenance of helium leak detection
equipment (which has its own vacuum
system). If neither a regular vacuum
technician nor a leak detector
maintenance technician is available, an
electronic technician or perhaps a
mechanical technician with some
electronics knowledge would be desirable.
Recognizing Abnormal Operation
of Vacuum Systems
There are really only two basic groups of
vacuum systems problems, though each
of these may be split into numerous
subheadings: (1) vacuum system and/or
mechanical problems and (2) automation
and/or electrical or electronic problems.
One of the most difficult and yet most
important questions to answer adequately
is just how well the machine would
perform under a given operational
condition — in other words, when a
machine is normal in operation and when
it is not.
For example, assume that all
automation and normal sequential
functions perform properly, but doubt
exists that the vacuum performance of the
machine is either normal or adequate
under the operational conditions existent.
It may be that the system is doing as well
as can be expected when its actual work
load, along with the time elapsed because
system cleaning and maintenance, are
considered. The best course of action in
this case is to discuss the present
operations and the previous operational
history of the vacuum machine with the
service engineer. If the information given
him is correct and complete, he or she
can evaluate the performance of the
machine in the light of his or her field
experience.
Performance of Vacuum System
during Starting Transients
It is possible that, with extensive auxiliary
equipment and heavy gas loads in the
vacuum system, pumping times greater
than normal may exist. It should also be
noted that the rated performance for
vacuum systems is for machines that are
kept running almost constantly and not
for equipment that has just been started
up after routine shutdown or recent
cleaning. When a machine has been
freshly cleaned or simply shut down for
some time, it may take 24 h or more
before routine operational pumping times
are obtained on a predictable basis.
240
Leak Testing
Discriminating between
Vacuum System
Contamination and Leaks
After it has been determined that the
vacuum performance of the system is
abnormal, it is important to decide just
what degree of malfunction is actually
present. This is important because the two
main problems will fall under the general
headings of system contamination and
system leaks. Whether or not a system is
leaking or is contaminated is sometimes
quite difficult to determine. However, if
the vacuum system has been operating
normally and has apparently slowly
degraded in performance to an
unacceptable but not catastrophic level, it
is probably subject to contamination
problems of one sort of another.
It is also necessary to consider any
recent work done on vacuum systems
because this, of course, could be a
potential cause of system leaks. However,
if vacuum performance has degraded
rather drastically, especially to the point
where only roughing level vacuum can be
obtained, a leak is almost certain and
troubleshooting procedures should be
oriented around that assumption.
Residual gas analysis indicating a high
nitrogen peak will often suggest a leak as
opposed to contamination.
The most difficult vacuum system
problems to solve are those where
degradation is definitely moderate by any
standard and could thus be caused by
either system contamination or system
leaks. If such appears to be the case, it is
highly advisable that a thorough mass
spectrometer leak detection test be
performed. This is, as a matter of fact, a
procedure that many use immediately on
any vacuum system where performance
levels have dropped to an unacceptable
figure. It is a desirable procedure, because
once leaks are eliminated as a source of
trouble the only problem left is
discovering and remedying the source of
system contamination.
Problems Caused by
Contamination within
Vacuum Systems
As previously mentioned, one of the
broad basic causes of poor vacuum
performance is system contamination. It
is also possible for the mechanical pump
oil to become contaminated, which in
itself can cause poor pumping
characteristics. Before disassembling or
cleaning an entire vacuum pump system,
one of the first things to check is the
condition of the pump oil. Immediately
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
flushing and refilling of mechanical pump
oil is called for if any indication of
discoloration, low operating level or
thinning out of the oil itself is evident.
Many unnecessary cleaning jobs have
been done because the mechanical pump
was not routinely flushed and filled first.
It should also be noted that even
though the roughing pressure may appear
normal, this may be misleading to the
extent that the mechanical pump may be
just able to hold this pressure with no
pumping capacity in reserve. Should this
be the case, normal roughing pressures will
be produced, but the moment a work leak
is encountered, system performance will
suffer. It never hurts to change the oil in
the roughing pump. Be sure to flush only
with specified roughing pump oil. Never
under any conditions use acetone or other
solvents in any mechanical pump.
Problems Caused by
Contamination of Cold
Traps in Vacuum Systems
If it is found that no performance
improvement is attained after servicing
the mechanical pump or pumps and
attempting another system pumpdown,
the next step before attempting complete
disassembling and cleaning of the vacuum
pump system is to follow the
maintenance manual procedure for
complete vacuum system shutdown.
Then remove, inspect and thoroughly
clean the cold traps, baffles and
cryopanels. After heavy use with dirty
work loads, deposits accumulating on
these cryopumps may reduce their ability
to freeze out moisture due to the
insulating effect of the previously trapped
compounds. They may also produce a
long term slow leak effect due to the
outgassing of the materials deposited on
their surface. This is why a vacuum
system left running without liquid
nitrogen after having been exposed for
some time to heavy work load will often
achieve substantial better vacuum when
left running over a weekend. Sooner or
later, the contamination on the cold
traps, baffles and cryopanels will complete
its outgassing and be pumped out of the
system. In extreme cases, however, actual
removal and cleaning of cold traps, baffles,
cryopanels and chamber interior will
restore system performance much quicker
than attempting to clean only the pumps.
Changing Oil in Diffusion
Pumps
A question that arises when the vacuum
pump system has been shut down and the
cold traps or baffles removed for cleaning
is whether or not to remove the diffusion
pump for cleaning and an oil change.
This may be a very difficult question to
answer. One should consider the degree of
system malfunction, the length of time
since the oil has been replaced and
whether or not the vacuum system was
ever inadvertently exposed to the
atmosphere during operation. This is
sometimes caused by improper operation
and a hand operated valve or by
accidental tripping of the wrong valve
when an automatic system is operated in
the manual mode. It should be noted that
a system may stand a great deal of abuse
in this particular area. However, if a
system has been in operation for six
months to a year and conditions have
been moderately adverse, it would be
considered good practice to change the
diffusion pump oil. If the old oil has been
cracked due to exposure to the
atmosphere, then the pump should be
cleaned before the new oil is added.
Preliminary Operation
Following Maintenance
Work on Vacuum Systems
After mechanical pumps have been
cleaned and flushed, their oil changed,
belt tension checked and adjusted, hose
connections routinely tightened and
checked, cold trap and baffles cleaned and
the diffusion pump cleaned and the oil
replaced, the system should then be put
through a normal startup and pumpdown
procedure and allowed to run for at least
24 h.
Performance checks should then be
made on the system. It is very likely at
this time that the system performance will
be close to original specifications. If the
diffusion pump oil was changed,
performance is likely to improve during
several initial days of operation as the
diffusion pump oil becomes conditioned.
This is a common occurrence in all
diffusion pump vacuum systems.
If performance does not improve after
the above procedures have been
accomplished and thorough leak testing
with a helium mass spectrometer leak
detector has revealed no system leaks, it is
then safe to conclude that cleaning of the
entire vacuum system is necessary. This,
of course, could have been done
immediately on noticing the first
malfunction symptoms. However, the
previous procedure is recommended
because total cleaning is frequently
unnecessary and takes a much longer time
to accomplish than the routine cleaning
described.
Leak Testing of Vacuum Systems
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
241
Detection and Repair of
Leaks in Vacuum Systems
The process of helium mass spectrometer
leak testing in high vacuum systems
involves procedures and considerations
described below. No differentiation is
made here between manual and
automatic operation because the basic
vacuum plumbing system is identical,
with the exception of hand operated
rather than air operated valves. There are
three general headings under which leaks
may be classified: (1) gross single or
cumulative leaks, (2) small single or
multiple leaks and (3) virtual leaks.
Causes and Detection of Single
Gross Leaks in Vacuum Systems
The single gross type of leak is usually one
wherein a sealing member is or has
become totally ineffective. This may occur
as the result of an inadvertently pinched
O-ring seal or improper welding. Often a
gross leak of any type is also defined as
one wherein the vacuum system cannot
be rough pumped to below 100 Pa (1 torr)
in the specified time for the pump system.
However, it is usually found that, if
roughing pumps cannot reduce pressure
to the 100 Pa (1 torr) range, a seriously
damaged seal will eventually be
discovered.
Testing for a very large single leak with
a throttled leak detector requires a slow
and thorough operation. If a leak is such
that pressure in the vacuum system only
reaches the 100 to 50 Pa (1 to 0.5 torr)
range, it may be easier to locate the leak
by the vacuum gage tracer gas technique.
It should be noted here that many
modern leak detectors have gross leak
testing capabilities. Refer to each
manufacturer’s specifications.
Causes and Detection of Gross
Cumulative Leaks in Vacuum
Systems
Gross cumulative leaks, usually defined as
several rather large leaks in vacuum
system, give rise to the same lack of
performance as that caused by a gross
single leak. All the same procedures apply
in dealing with gross cumulative leaks
with the exception that, although large
cumulatively, they may be too small
individually to respond to the thermal
conductivity gage spray leak test. If it is
suspected that several leaks are causing
the system failure (and this may indeed
be the case, particularly if the system has
been cleaned and reassembled by
inexperience personnel), it may be
advisable to engage a service engineer for
assistance in remedying the problem. This
242
Leak Testing
is suggested because gross cumulative
leaks in any vacuum system usually
appear only during construction, after a
system has been subjected to physical
punishment or after inexperienced
personnel have attempted the
disassembly, cleaning and refitting of the
vacuum plumbing. Inexperience in
making vacuum seals may cause sealing
surfaces to be damaged. Also, careless
handling of parts, such as the
overextension of brass bellows while it is
removed from a bellows sealed valve, may
cause rupture and should always be
considered as a possible cause of gross
leakage in a vacuum system.
Remember when testing for leaks due
to ruptured bellows assemblies that a
bellows will give no indication of a leak
when the valve is closed unless leak tests
are made through the vent on the
atmospheric side of the valve to which
the bellows is still exposed. Test
possibilities may be found by examining
drawings of bellows stem sealed valves.
One may find that the leak can be located
by using a leak detector connected to the
pump valve or, in some cases, the vent
valve. Judiciously opening and closing the
suspected leaky valve while leak testing
the dysfunctional bellows may permit its
identification as the source of leakage.
Causes and Detection of Small
Single or Multiple Leaks in
Vacuum Systems
Small single or multiple leaks are readily
located with a helium mass spectrometer
leak detector. These types of leaks may
allow a vacuum system to be evacuated at
least into the low pascal range and usually
into the high vacuum range. Perhaps the
ultimate vacuum system pressure would
be only about 0.5 Pa (3.75 mtorr). Under
these conditions, a helium mass
spectrometer leak detector properly
connected to the vacuum system in
question will quickly enable these small
leaks to be detected. All suspected areas
are helium tracer probed or bagged
methodically in sequence while using a
suitable leak testing procedure.
Causes of small single or multiple leaks
are most often: (1) flanges that have been
improperly tightened; (2) O-rings that
have simply aged and taken a set;
(3) undamaged O-rings that are
improperly seated; (4) electrical
feed-through seals; (5) tiny cracks in
ionization gage tubes; (6) improperly
fitted gage tubes; (7) poor fitting and/or
seating of gaskets; and (8) weld joints that
leak after repairs or on completion of new
systems.
Any or all of these may contribute to
small single or multiple leaks.
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
PART 5. Equipment and Techniques for
Measuring Pressure in Vacuum Systems
Introduction to Vacuum
Gages
As important as the production of
vacuum is the ability to gage its results
through pressure measurement. Various
types of commercial gages are available
that cover the pressure range from
atmospheric pressure to less than 10 µPa
(100 ntorr). In the high pressure region,
gages are used that depend on the actual
force exerted by a gas. At low pressures,
some specific property of gases (such as
thermal conductivity or ability to become
ionized) is used as the basis for measuring
pressures.
Gages are generally calibrated in
pressure units such as millipascal or
micropascal (or the older units of torr or
bar). The various types of common
vacuum gages may be summarized as
follows.
1. Pneumatic force gages depend on the
actual force exerted by the gas.
Examples are mercury and oil
manometers, McLeod gages, Bourdon
gages and diaphragm gages.
2. Thermal conductivity gages depend on
the change of the thermal
conductivity of a gas with change of
pressure. The most common examples
are the Pirani and thermocouple gages.
3. Ionization gages depend on the
measurement of electrical current
resulting from ionization of gas.
Examples include thermionic
ionization gages (Bayard-Alpert), cold
cathode gages (penning or Philips) and
alphatron gages.
Bourdon and Diaphragm
Vacuum Gages
The Bourdon and diaphragm gages are
mechanical gages that are used primarily
for giving an indication that a vacuum
system is actually below atmospheric
pressure. Most of these gages indicate
negative gage pressure from atmospheric
pressure down to their lower pressure
limit in the low pascal range (a fraction of
a torr). They can be constructed of
noncorrosive materials to make it possible
to use them in the presence of corrosive
gases and vapors. Because they work on
the basis of the force exerted by a gas,
they measure the total pressure of a
mixture of gases and vapors.
Operation of Bourdon Vacuum
Gage
Bourdon gages, shown in Fig. 21a, make
use of a tube that is sealed off at one end
with the other end leading to the
connection to the vacuum system. The
tube is usually of elliptical cross section
and is bent into an arc. A change of
pressure inside the tube makes it change
its curvature. This change is transmitted
through a series of levers and gears to a
needle that gives a reading of the pressure
on a circular scale behind the needle. As
shown in Fig. 21b, the calibration of the
scale in pascal absolute should ideally
have 100 000 on top center, 0 at left
bottom and 200 000 at right bottom. A
few gages in North America are still based
on inch of mercury, from 0 to 30 in. Hg,
where 0 represents atmospheric pressure
and 30 represents a good vacuum.
Actually, the accuracies of most Bourdon
gages may not be sufficient to read a good
vacuum: the smallest reading is about
1 kPa (0.01 atm). However, these gages are
occasionally still used to indicate the
condition of a vacuum system.
Operation of Diaphragm Vacuum
Gage
The operation of the diaphragm gage
shown in Fig. 21c is based on transferring
the distortion of the diaphragm to a scale
reading. Diaphragm distortion is caused
by a pressure differential across it. The
scale may be calibrated in kilopascal, in
torr or in inch of mercury.
Operation of Liquid Level
Manometers (McLeod
Gages)
Before 1981, the gage used most
commonly as a comparison calibration
standard by the National Institute of
Standards and Technology and industry
was the McLeod gage, a mercury
barometer. It has since been replaced by
the spinning rotor gage and accepted by
the National Institute of Standards and
Technology as the primary standard. As a
Leak Testing of Vacuum Systems
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
243
FIGURE 21. Principles of operation of
mechanical vacuum gages: (a) elements of
Bourdon gage; (b) external appearance of
Bourdon gage; (c) elements of diaphragm
gage; (d) older English combination gages
with inch of mercury calibration on the left
and pound per square inch on the right.
(a)
Needle
Lower than
atmosphere
Scale
Higher than
atmosphere
Determining Gas Pressure from
McLeod Gage Reading
Elliptically
shaped
tube
Closed
Lever and gears
To vacuum
(b)
80
(1 atm)
100
120
60
140
160
40
kPa
20
180
200
0
Scale
(c)
result, the following description of the
McLeod gage will be abbreviated but
sufficient to understand it.
The principle is based on the
application of Boyle’s law and is quite
simple. A known volume of gas, at the
pressure that is to be measured, is trapped
and compressed by a known ratio to a
new pressure that may be determined. By
inserting the known values (original
volume, final pressure and final volume)
into the Boyle’s law formula (PiVi = PfVf),
the original pressure of the gas may be
computed.
Linkage
Needle
Reference
vacuum
Diaphragm
0 kPa
Atmospheric
pressure,
(100 kPa)
To vacuum
(d)
The gage is operated by raising the
mercury above the gage head cutoff point
indicated in Fig. 22a. A sample of the gas
to be measured is trapped by rising
mercury in the bulb volume between the
cutoff point and the top of the closed
capillary tube. This volume may be called
Vi and is determined by the manufacturer
when the gage is being fabricated. The
mercury level is raised until the level in
open capillary B is directly opposite the
top of the closed capillary tube A. The
mercury level is raised until h = h’. Raising
the mercury level has compressed the
sample volume of gas in the closed
capillary so that it occupies the tube
length, h. The sample has now been
compressed to a new volume Vf equal to
the cross sectional area of the capillary
tube times the height h. The head of
mercury, which is compressing it to this
volume, is also h’ = h. Applying Boyle’s
law, Eq. 28, it follows that:
(28)
=
Pf Vf
where Pi is pressure of gas sample to be
measured (unknown); Vi is bulb volume
(known); Pf is final pressure of compressed
gas sample which is indicated by the
height of the mercury column, h’ = h; and
Vf is volume of compressed gas sample
which equals gas column height h
multiplied by the cross sectional area a of
the closed capillary column. Inserting
known values in Eq. 28 yields Eq. 29:
(29)
Pressure
Pi Vi
Pi Vi
=
h (a h )
=
a h2
0
–10
5
–20
10
–30
in. Hg
244
Leak Testing
15
lbf·in.–2
Limitations of McLeod
Gage Measurements
The McLeod gage does not measure the
pressures of condensables in the vacuum
system. On the other hand, it is equally
sensitive to all gases that follow Boyle’s
law. Its biggest disadvantage is that it has
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
a discontinuous gage reading; continuous
readings of pressure variation in a system
are not obtainable with a McLeod gage.
Operation of Spinning
Rotor Gage
The spinning rotor gage (Fig. 23) has been
accepted by the National Institute of
Standards and Technology as a transfer
standard gage. This is possible because the
principle on which the gage works can be
FIGURE 22. Operating principle of the
McLeod gage: (a) head arrangement;
(b) quadratic scale measurement system
(h’ = h); (c) linear scale measurement system.
(a)
To vacuum
Open capillary B
Side arm
Closed capillary A
related through calculation to basic laws
of physics. Its name says exactly what it is
— a spinning rotor. Several manufacturers
produce them for use in metrology
laboratories or for industrial applications
where higher accuracy is needed without
a mercury manometer and its toxicity
related hazards.
A magnetized ball is magnetically
suspended in a small chamber to
eliminate all sources of friction except air
friction. It is made to spin or rotate while
suspended. If there are gases present in
the chamber, the ball will slow down due
to the impacts from molecules in the
chamber. The rate at which it slows down
is directly proportional to the gas pressure
(number of impacts). All that needs to be
done then is to very accurately measure
the rate at which the ball slows down and
calculate the pressure as a result. This is
done by measuring the frequency of the
magnetic pulses induced in the pickup
coils. The calculation is, of course, done
electronically by the attached control
unit.
One manufacturer of this gage states an
accuracy of 1 percent of the reading
±4 µPa (30 ntorr) between 10 µPa to 1 Pa
(70 µtorr to 10 mtorr). Although you will
not be using this gage as a routing
pressure gage, your system gages may be
calibrated using the spinning rotor gage.
Bulb
Cut-off
FIGURE 23. Spinning rotor gage.
Tube to reservoir
(b)
Vertical
magnetization
of ball
A
N
Reference line
h’
h
Permanent
magnet
Vertical
stabilization coil
S
B
Pickup
coil
(c)
Pickup coil
Lateral
magnetization
of ball
h
Vacuum tube
h0
Ball
N
Vertical
stabilization
coil
Permanent
magnet
Reference line
S
End view
cross section
Leak Testing of Vacuum Systems
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
245
exposed to the gas whose pressure is to be
measured. For absolute pressure
measurement, the other (reference) side
contains an electrode assembly placed in a
sealed high vacuum reference cavity.
Because the electrodes in the absolute
pressure gage are not exposed to the gases
being measured, this gage is not affected
by oil or water vapors or by corrosive or
other chemically active process gases.
The diaphragm deflects with changing
pressure force per unit area —
independent of the composition of the
measured gas. This causes a capacitance
Operation of Capacitance
Manometer
The capacitance manometer (Fig. 24) is
another pressure gage that can be used in
the rough vacuum range. It is capable of
measuring the absolute pressure or
relative pressure, depending on the gage
model used. It does respond to the total
pressure. It is not sensitive to changes in
gas mixture as are many other gages.
The sensing unit contains a tensioned
metal diaphragm, one side of which is
FIGURE 24. Manometer gage: (a) schematic of electronic system; (b) differential setting; (c) absolute setting; (d) components.
(a)
Output
connector
0 to 10 V
0 to 10 V
Amplifiers
(alternating current)
Amplifier
(direct current)
Demodulator
Preamplifier
Sensor
±58 V supply
Oscillator
10 kHz
± 15 V supply
(b)
Electrodes
PR
D
Px ← P
Differential
(c)
Px ← P
Evacuated and sealed
Absolute
(d)
PR port (differential only)
Capacitor electrode
Sensor body and diaphragm assembly
Getter assembly (absolute only)
Electrode
connections
246
Leak Testing
Px port
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
change between the diaphragm and the
adjacent electrode assembly. The
capacitance change is sensed in an
oscillator circuit and converted to a
frequency change proportional to the
diaphragm deflection.
This frequency change, in turn, is
converted in the unit to be displayed as
the pressure reading. The sensor unit may
be constructed of materials such as nickel
base alloy and stainless steel, allowing the
gage to be used with corrosive gases.
This gage is sufficiently accurate (about
1 percent of reading) and precise that one
can worry about the effect of temperature
changes (Charles’ law) on the pressure
readings. The sensor head may be placed
in a constant temperature oven as a result.
This gage is often used as a flow controller
because of its fast response (milliseconds)
to pressure changes. If you desire to use a
capacitance manometer over a wide
range, you may need several units. The
gage is constructed to read over three or
four orders of magnitude. If you wish to
read from atmosphere (760 torr) into the
high vacuum range (10 µtorr), that is
seven orders of magnitude. Therefore, you
need several different gage units. These
gages can be constructed so that pressures
from 105 to 10–5 torr may be sensed, but
any particular gage is limited to about
four orders of magnitude of that range.
Below 0.1 Pa (1 mtorr) the accuracy falls
dramatically.
The capacitance manometer may
receive more maintenance than many
gages because of its ability to read
accurate and precise pressure values. It
may periodically be taken to the
calibration lab for a check against some
standard gage. When it is used in dirty or
corrosive gas systems, the sensing side of
the gage head may be flushed with an
appropriate solvent.
Overpressuring the gage (20 percent
over full scale) may shift the reading or
permanently damage it. An isolation valve
is often used to prevent this.
The rate of heat transfer in a low pressure
gas depends in a complex manner on the
specific heats, molecular weight,
temperature and pressure of a particular
gas. Under suitable conditions it can thus
be used as an indication of the pressure.
The useful pressure range of thermal
conductivity gages extends from 270 Pa
(2 torr) to about 0.1 Pa (1 mtorr), where
the rate of heat transferred by radiation
begins to predominate over the rate of
heat transferred by conduction in the gas.
The two most common types of thermal
conductivity gages are the Pirani gage and
the thermocouple gage. In both gages,
conductivity changes of a gas cause a
variation in the heat losses from an
electrically heated filament. This
temperature change is measured by means
of a thermocouple in the thermocouple
gage. A bridge circuit measures the change
of electrical resistance of the heated
filament in the Pirani gage.
Construction and Operation of
Thermocouple Vacuum Gage
Figure 26 shows a simplified schematic of
a thermocouple gage circuit. A
thermojunction of two thin dissimilar
metals are connected to the midpoint of a
tungsten heater wire that is supported
inside a metal envelope attached to the
vacuum system. A constant current of the
order of 30 mA is passed through the
heater wire. The thermal electromotive
force developed across the thermocouple
wires is of the order of 10 to mV and may
be read on a simple meter. The
temperature attained by the thermocouple
FIGURE 25. Principle of the thermal conductivity (Pirani)
gage. Thermal losses from the electrically heated resistance
wire vary with heat conduction by gas molecules. Heat
losses are reduced as gas pressure is lowered.
To vacuum
Measuring Pressure in
Vacuum Systems with
Thermal Conductivity
Gages
Heat transfer through a gas is related to
the molecular density of the gas between
surfaces across which a temperature
difference exists. As gas molecules are
removed from a system, the amount of
heat transferred by conduction in the gas
is also reduced. Finally, at a sufficiently
low pressure, heat transfer within a
thermal conductivity gage occurs by
radiation and convection losses, while
conduction effects are negligible (Fig. 25).
Conduction
through gas
molecules
Radiation to
surroundings
Heated wire
Heat loss
through
conduction
Leak Testing of Vacuum Systems
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
247
depends on the conductivity of the gas
surrounding the junction and thus on the
pressure. The gage is calibrated to read on
a logarithmic scale whose range may be
extended upward by incorporating the
convection principle with some reduction
in accuracy.
The thermocouple gage, though not as
accurate as the Pirani in vacuums near
10 mPa (or 0.1 mtorr), is more than
adequate for forepressure measurements.
Because of its simplified circuit, it is only
about half as expensive as a Pirani gage
and can be easily packaged into
multistation vacuum leak testing
instruments.
Advantages and Limitations of
Thermocouple Vacuum Gages
The thermocouple gage has the virtue
of simplicity and the disadvantage of a
nonlinear scale. The calibration of the
thermocouple gage may be changed by
changing the heater current. A low value
of heater current and a sensitive meter in
the thermocouple spread the scale at low
pressures. High current and a less sensitive
meter spread the scale at higher pressures.
The advantages of the thermal
conductivity gages for industrial
application are numerous. They respond
to vapors, read continuously and
remotely, need not be fragile or bulky and
may be used in automatic control
systems. Their selective response to
hydrogen and helium makes them useful
for leak hunting. No damage is done to
these gages if the vacuum system is
exposed to atmospheric pressure while
they are on.
Circuit and Operating Principles of
the Pirani Vacuum Gage
Pirani gages use a Wheatstone bridge
circuit, as shown in Fig. 27, which serves
to heat a filament and to balance its
resistance against a standard resistor
sealed off in high vacuum. A change of
pressure causes a change of filament
temperature and, consequently, of the
filament resistance, thus unbalancing the
bridge. The pressure can then be
measured in terms of the unbalanced
voltage. Alternatively, the power required
to maintain the filament temperature at a
constant level is a measure of pressure.
The temperature in this case is kept
constant by means of feedback circuit.
The sensitivity of a Pirani gage
decreases rapidly as the pressure is
increased, owing to the fact that collisions
between gas molecules become more
frequent and that the thermal
conductivity tends to become
independent of the pressure. In the usual
Pirani gage, a dummy tube (compensator)
just like the one connected to the vacuum
is used for one arm of the bridge. This
tube is highly exhausted and sealed off.
The two tubes are mounted together so
that they will have the same ambient
temperature. The bridge is balanced while
the gage tube is under vacuum. The
unbalanced current of the bridge is then
taken as an index of pressure.
More recent digital readout Pirani gage
designs incorporate compensating
networks within the Wheatstone bridge to
FIGURE 26. Simplified thermocouple gage circuit.
FIGURE 27. Pirani gage circuit.
To vacuum system
To vacuum
Standard resistor
sealed in a
dummy tube
Seal
Thermocouple
Gage
Meter
calibrated
in pressure
units
Meter
Heated filament
Seal
Power
supply
Electrical power supply
Heater current
adjust
248
Leak Testing
Meter
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
produce fairly accurate absolute pressure
readings from atmosphere to 0.1 or
0.01 Pa (1.0 or 0.1 mtorr).
Creation of Ions in
Ionization Gages Used for
Measuring Vacuum
A neutral particle (atom or molecule)
contains the same number of positively
charged protons in the nucleus as
negatively charged electrons in the orbits
around the nucleus. By detaching one of
the electrons from a neutral particle, a
positive molecular or atomic ion is
produced. The process is called ionization.
This positive ion will be influenced by the
same electric and magnetic forces that
influence an electron, but in the opposite
direction. For example, a negatively
charged plate will attract a positive ion.
Ionization is fairly easily accomplished
by electron bombardment. Electrons of
sufficient energy, directed at a neutral
particle, cause an energy transfer whereby
the orbital electron attains sufficient
energy to overcome the atomic forces that
bond it to the nucleus. The orbital
electron leaves its orbit as a free electron,
leaving behind a positively charged ion.
The ability of a gas to become ionized is
the basis of ionization gages.
Types of Ionization Gages
Used to Measure Vacuum
The different types of ionization gages
vary in the manner of forming positive
ions and in the manner of collecting
them. all require calibration, although
variation in sensitivity within a particular
model is not great. The two ionization
gages most commonly used are (1) the
cold cathode or discharge gage (Philips
gage) and (2) the thermionic ionization
gage (Bayard-Alpert gage).
Of the several types of ionization gages,
all have the common feature of measuring
an ionization current that is proportional,
for any one gas, to the molecular
concentration. However, the probability
of ionization of a molecule by
bombardment by a charged particle is
almost independent of the velocity of the
molecule. Thus, the gage actually operates
by measuring the molecular concentration
in its electrode region rather than the
pressure there.
Cold Cathode Vacuum
Gages
The cold cathode type of vacuum gage is
also known as the Philips discharge gage
or Penning gage. In the cold cathode gage
(Fig. 28), electrons are drawn from the
two plate type cathodes by the
application of a high voltage and are
attracted to the positive anode. The path
of the electrons from cathode to anode is
made several hundred times longer by
arranging a magnetic field across the tube
in the direction shown. The path now
traveled by the electrons is a helix rather
than a straight line. The increase length
results in a proportional increase in the
probability that an electron will ionize the
molecules of residual gas by collision. An
ionization current is produced that is
several times greater than that which
would be produced if no magnetic field
were present. Actually, the total discharge
current (the sum of the electron current
from the cathode and the positive ion
current to the cathode) is used as a
measure of pressure in the system. No
amplification of the discharge current is
necessary and it may be fed directly to a
pressure indicating microammeter that
responds to the net current.
Performance Characteristics of
Cold Cathode Ionization Gages
The range of cold cathode gage pressure
measurements extends from 100 Pa to
10 µPa (0.5 torr to 0.1 µtorr). Because of
its simplified circuit, this type of
ionization gage is relatively inexpensive.
Because the resistance changes with
pressure, the ionization current output is
nonlinear. The most accurate readings are
obtained between 100 and 0.1 mPa (1 torr
to 1 µtorr) where they can be used for fine
pressure measurements. the cold cathode
gage is not subject to sensing tube failures
as a result of exposure to high pressures or
a sudden loss of vacuum. Because of the
FIGURE 28. Principle of cold cathode discharge gage.
Transverse magnetic field
–
–
–
+
+
Anode (+)
Cathodes (–)
+
–
+
–
–
Leak Testing of Vacuum Systems
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
249
heavy type of construction, the tube is
not easily degassed. It is more readily
contaminated owing to the high rate of
ionization existing within the tube.
Therefore, cold cathode gages should not
be used for forepressure measurements.
Design and Construction of Cold
Cathode Ionization Gages
The most common commercial cold
cathode discharge gages do not use
separate cathode plates. The trend has
been instead toward all-metal
construction with the inside wall of the
tube acting as the cathode. The anode is
usually in the shape of a ring, but also
may be round, square or rectangular
(Fig. 29). In some cases, use is made of a
wire loop anode sufficiently heavy to
prevent vibration and sagging. A compact,
high strength alloy magnet is used.
Usually, the magnet and gage tube are
made as a single unit. Stainless steel,
aluminum and nickel plated copper are
used in commercial gages for the tube
body (cathode). Theoretically, the cathode
material should not sputter readily so that
it will not produce a conducting layer on
the insulator through which the anode is
connected.
Principle of Operation of
Thermionic Ionization
Gages
The electrons usually do not hit the grid
structure when they first reach it, but
oscillate through it several times before
being collected. An emission regulation
circuit is used to keep the electron current
at a steady value. Positive ions formed
between the gird structure and an outer,
cylindrical collector electrode are attracted
toward the collector maintained at about
–20 V. This positive ion current, flowing
to the ion collector electrode, is
FIGURE 30. Hot filament ionization gage: (a) principle;
(b) construction; (c) simplified electrical circuit.
(a)
Filament cathode
–
+
+
–
+
Ions
+
+
Collector (plate)
Electrons
Grid
(b)
The hot wire ionization gages is most
widely used for measuring absolute
pressure below 100 µPa (1 µtorr). Its
operation depends on ionization of a gas
with electrons emitted from a heated
filament. The ions thus produced are
collected and the resulting current
measured. The most common version of
the gage (Fig. 30) uses a tungsten or thoria
coated iridium hairpin filament to emit
an electron current of about 5 mA. The
electrons are accelerated outward toward a
cylindrical grid operated at about +150 V.
Tube envelope
Plate
To vacuum
Grid
Filament
FIGURE 29. Commercial cold cathode gage.
Seals
Anode shield
Magnet pole piece
Fluorocarbon
resin
O-ring
(c)
Plate
Grid
Filament
Anode loop
Gage body
(cathode)
250
Leak Testing
M
Meter calibrated
in pressure units
Anode flange
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
proportional to gas density over a wide
pressure range.
Performance Characteristics of
Thermionic Ionization Pressure
Gages
The lower pressure limit for the gage
configuration of Fig. 30 is 1 µPa (10 ntorr)
The limitation is due to an X-ray effect
that produces a constant residual collector
current irrespective of pressure. Electrons
arriving at the positive grid produce x-rays
that irradiate the negative ion collector
and release from its surface
photoelectrons that are attracted to the
positive electrode. The current of
photoelectrons leaving the ion collector is
indistinguishable from a current of
positive ions arriving, down to pressures
of 1 µPa (10 ntorr). The photoelectron
current is roughly proportional to the
surface area of the ion collector and
surface area of the grid.
Operating Principle of
Bayard-Alpert Gage for
Pressures down to 1 nPa
(75 ptorr)
For accuracy in reading pressures below
1 mPa, the constant residual collector
current must be reduced to as low a level
as possible. The Bayard-Alpert
modification of the thermionic ionization
gage accomplishes this by inverting the
structure as shown in Fig. 31. The
filament is outside the cylindrical grid,
which acts as a positive potential to
collect the electrons. The ion collector is
at a negative potential and consists of a
fine wire suspended centrally within the
grid.
Because the area of the ion collector
exposed to radiation from the grid is
about 100 times smaller than that in the
conventional gage, the production of
photoelectrons and, therefore, of the
residual constant background current is
reduced proportionally. This makes it
possible to measure ion currents
corresponding to pressure of the order of
10 nPa (0.1 ntorr). Most of the X-rays are
absorbed in the Bayard-Alpert gage by the
glass envelope. However, to measure low
pressure, it is necessary to thoroughly
outgas the tube. Outgassing is usually
accomplished by electrically heating the
grid.
(2) high-frequency oscillations and
(3) decomposition of gas. Gage pumping
action is a chemical as well as an electrical
phenomenon. Chemical pumping at
8 mA electron current and 150 V electron
energy is less than 2 L·s–1 (4.25 ft3·min–1)
for nitrogen. This pumping action causes
the gage to indicate lower system pressure
than actually exists. High frequency
oscillation in the gage may cause a
buildup of potential as much as –150 V
on the glass walls. This may have a
serious effect on the gage sensitivity,
especially between 100 and 10 mPa (1.0
and 0.1 mtorr). Some manufacturers coat
the inside of the glass walls with a
metallic film to remove this potential,
thus increasing its accuracy.
Gas decomposition is encountered
when the tungsten filament is operated at
2000 K (3140 °F). The most effective way
to reduce this problem is by reducing the
filament temperature. Thoria coated
iridium filaments have been successfully
used, providing high emission at
relatively low temperature.
Calibration of Thermionic
Ionization Gages for
Different Gases
A thermionic ionization gage has different
sensitivities for different gases. In reality,
the gage measures molecular
concentrations rather than true pressures.
A gage measuring the pressures of two gas
samples at different temperatures, but
having the same pressure for both
samples though the higher temperature
sample really has a higher pressure.
FIGURE 31. Bayard-Alpert gage.
Electrometer
To vacuum
Ion
collector
Degassing coil
Filament
Power supply
Performance Characteristics of
Bayard-Alpert Vacuum Gages
Major sources of error in pressure
measurement with the Bayard-Alpert
gages are (1) pumping action of the gage,
To filament supply
Leak Testing of Vacuum Systems
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
251
Leak Testing with Bayard-Alpert
Electronic Gage
Experience indicates that the
Bayard-Alpert hot filament pressure gage,
when used as an electronic leak detector
on small volume systems, provides
solutions to some of the problems of
system leak detection encountered with
the helium mass spectrometer. Unlike the
spectrometer, the electronic leak detector
uses a system’s own vacuum pump, which
TABLE 3. Calibration of Bayard-Alpert ionization gages
for different gases. Multiply ion gage reading by factor
shown for correct pressure. To get sensitivity in µA·Pa–1,
divide 750 by gage factor (or µA per µtorr, divide 100 by
gage factor).
Sensitivity
_______________________
Gas or Vapor
Air
Argon
Carbon dioxide
Carbon monoxide
Helium
Hydrocarbon pump oil
Hydrogen
Krypton
Mercury
Neon
Nitrogen
Oxygen
Silicone pump oil
Water
Xenon
252
Leak Testing
Gage Factor
1.10
0.84
0.73
0.94
6.20
0.20
2.00
0.53
0.29
0.42
1.00
1.18
0.37
1.12
0.37
µA·Pa–1
682
892
1030
800
121
3750
375
1420
2580
1790
750
634
2030
670
2030
(µA·µtorr –1)
(91)
(119)
(137)
(106.5)
(16.4)
(500)
(50)
(189)
(344)
(238)
(100)
(84.5)
(270)
(89.3)
(270)
is proportioned to the size of the system
on which it is used. To operate, the
detector need only be connected to the
controller on the selected detecting device
(either pump or gage) and an electrical
outlet.
With the electronic detector, a small
volume system can be leak tested at
virtually any pressure at which it based
out. When a leak is found, it can often be
temporarily closed with plastic sealant
and use of the system can continue until
a permanent repair can be effected, thus
avoiding wasted runs and down time.
With the electronic detector, response to a
leak is extremely rapid, regardless of the
size of the system. Furthermore, cleanup
time (that time required, once the tracer
gas has been removed from the leak, for
the background of tracer gas to dissipate,
restoring a good signal-to-noise ratio) is
remarkably short. Finally, the detector
does not require liquid nitrogen and does
not restrict the user to helium as a tracer
gas. Although oxygen and argon give the
greatest sensitivity, many other gases can
be used effectively.
On the other hand, the measurement
of a leak with the electronic detector
presents one problem not encountered
with the helium mass spectrometer.
Unlike the spectrometer, the electronic
FIGURE 32. Actual pressure versus indicated gage pressure for
Bayard-Alpert gage.
Actual pressure, Pa (lbf·in.–2 × 1.45)
The actual pressure of a particular gas is
dependent in a complex fashion on the
mass of the gas molecule and its
ionization energy. These factors, though,
are constant, so that a gage calibrated for
nitrogen may accurately read the pressure
of other gases by simply multiplying the
indicated pressure by a constant factor.
As an example, consider a gage that is
calibrated for nitrogen and reads 0.1 µPa.
If the system is evacuated and backfilled
with nitrogen, then it can be assumed
that the total indicated pressure is almost
completely due to nitrogen and therefore
an actual pressure of 0.1 µPa (1 ntorr)
exists. If, instead, the system was
backfilled with helium, the total indicated
pressure would be due almost entirely to
helium and the actual pressure would be
6.2 × 0.1 mPa (6.2 × 1 µtorr) — 6.2 is the
correct multiplication factor for helium.
Table 3 lists correction factors for
different gases. Figure 32 is a graph of
actual pressure versus indicated pressure
for three gases, air, helium and argon, for
a Bayard-Alpert gage.
100
(10–4)
10–1
(10–5)
10–2
(10–6)
10–3
(10–7)
10–4
(10–8)
10–5
(10–9)
10–6
(10–10)
10–7
(10–11)
10–8
(10–12)
10–9
(10–13)
10–9
10–8
10–7
10–6
10–5
(10–13) (10–12)(10–11) (10–10) (10–9)
10–4
10–3
10–2
10–1
(10–8)
(10–7)
(10–6)
(10–5)
Gage reading, Pa (lbf ·in.–2 × 1.45)
Legend
=
=
=
=
Helium
Air
Nitrogen
Argon
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
detector cannot be calibrated in absolute
units. This does not mean that the
electronic detector necessarily has a low
sensitivity, but rather that its sensitivity
varies with the size of the total gas load of
the system on which it is used.
The electronic leak detector has the
advantage that it is almost impossible for
an operator to inadvertently damage it,
the system on which it is being used or
any instruments on that system. However,
getting the best performance out of the
instrument requires a reasonable amount
of operator skill and experience.
and for all work with components. For
the engineer interested in low pressures
on small to medium systems, however,
portability, ease of operation and low
price (about one tenth the price of the
helium mass spectrometer) make the
electronic detector an extremely valuable
tool.
Sensitivity Limitations of
Bayard-Alpert Gage Used As a
Leak Detector
As with any electronic device, the
sensitivity of a Bayard-Alpert pressure
gage is limited by the signal-to-noise ratio.
The noise encountered comes from many
different sources and is found to cover a
broad frequency spectrum. The higher
frequency noise sources are often the ion
gage connections and the amplifier itself.
Good connections and shielding should
be maintained throughout this part of the
ion gage circuit. Effects should be made to
reduce the flow of cooling air currents
about the gage tube and the movement of
the collector cable during leak detection.
The amplifier and, particularly, the
filament emission regulator circuit should
be working correctly to avoid variations in
collector current. In the case of the ion
pump, pressure changes due to gas bursts
or leakage current in the pump can be a
source of fluctuation. The pump history
may show a cause for these effects and
they may be cured by bakeout or high
potential electrical testing in certain cases.
Noise originating in the alternating
current line should be largely eliminated
by the filtering system in the leak
detector.
Very low frequency noise or drift,
having a time constant in the order of
minutes, may be caused by a number of
conditions. For instance, the system gas
load may be changing, as is the case
during pumpdowns or when the system is
subject to thermal drift. In such cases, it is
proper to wait until the system has based
out and/or the thermal drift has been
eliminated before leak testing. However,
electronic detectors are normally supplied
with an output connection to which a
strip chart recorder can be attached. The
deflection on the strip chart is of a
definite and characteristic form, which
allows it to be separated with reasonable
ease from the background noise.
Obviously, the electronic leak detector
is not the answer to all leak detection
problems. It is impractical for work that
requires absolute measurements of leaks
Leak Testing of Vacuum Systems
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
253
PART 6. Techniques for Detection of Large
Leaks in Operating Vacuum Systems
Problems in Locating
Gross Leaks in the Coarse
Vacuum Range
Large leaks can be the most difficult and
exasperating ones to find in vacuum
systems. Most of the sensitive techniques
and equipment developed for leak
detection in vacuum systems are
inapplicable at the pressures attainable by
vacuum system pumps when large leaks
are present (100 to 0.1 kPa or 760 to
1 torr). Consequently, large leaks usually
are sought by one or another of a number
of relatively crude techniques. Some of
these tests are based on pressure testing or
bubble leak testing techniques.
Design of Vacuum Systems
for Convenience of Leak
Testing during Operation
Because almost every (if not every)
vacuum system will leak at one time or
another during its lifetime, it is well to
give some thought to the problem of ease
of leak testing during the design of a
vacuum system. A great amount of time
can be wasted if poor leak hunting
techniques must be used because it is too
difficult or impossible to use a better
technique on the existing system. The
lack of forethought in this matter is all
the more deplorable because improving
vacuum system design to get better leak
hunting efficiency usually requires only
simple and relatively inexpensive
measures, such as proper location of a
valve or gage that will be in the system
anyway.
It should be possible to isolate the
roughing pumps from the system with a
valve that can also be used to throttle the
pumping speed of these pumps. A thermal
conductivity gage should be placed in the
fore vacuum line between this valve and
the diffusion, turbomolecular or ion
pump, for use in rate-of-rise
measurements as well as to monitor the
fore pressure. A stub into the foreline
should also be inserted at this point for
connection of a vacuum leak detector. The
stub should have a valve and connection
fitting (a flange that mates with the leak
detector, a quick disconnect fitting or the
254
Leak Testing
like). This connection for a leak detector
should be of fairly high conductance in
order that response time not be impaired.
If only the basic version of leak detector
(without roughing pump) is available, it
will help to have another valve between
the stub and the turbomolecular pump or
diffusion pump, so that the roughing
pump for the system can be used to
evacuate the line to the leak detector.
If possible, it is desirable to have a
valve between the high vacuum chamber
and the diffusion pumps. It need not be
possible to throttle the pump with this
valve, its main purpose being to isolate
the chamber for either isolation or
rate-of-rise tests. The chamber itself
should have one or more ionization gages
(even if an ion pump is used) in addition
to any ultrahigh vacuum gage that may
be used.
Leakage Rates Tolerable in
Operating Vacuum
Systems
Leaks can be tolerated in an operating
vacuum system if the mass flow rate of
the leak plus any outgassing load does not
exceed the capacity of the pump at the
operating pressure. For example, a system
that must be maintained at 10 µPa
(0.1 µtorr) with a 0.1 m3·s–1
(200 ft3·min–1) pump can handle
10 × 10–6 × 100 × 10–3 = 10–6 Pa·m3·s–1
(10–5 std cm3·s–1) of gas. So long as the
sum of all leakage and outgassing is less
than this value, the vacuum system
operating pressure of 10 µPa (0.1 µtorr)
will be obtained and there is no need to
search for leaks smaller than about
10–7 Pa·m3·s–1 (10–6 std cm3·s–1) in this
system. If there are leaks larger than can
be handled by the vacuum pumps, one of
the techniques to be described can be
used to locate the leak. In most cases the
actual value of the leakage rate is not
desired, although it can be obtained by
using calibrated leaks with the leak
detector on smaller volume systems or by
using system calibrated leaks on very large
volume systems.
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
When a vacuum system fails to reach the
ultimate pressure that has previously been
obtained with the system or which is
expected for other reasons, air leakage
into the vacuum system is to be
suspected. However, a vacuum system
that takes an unusually long time to reach
its ultimate pressure (and for practical
purposes fails to reach this pressure) often
has internal sources of gas and vapor
rather than leakage from outside the
system. Before embarking on extensive
leak hunting, the possibility of internal
gas sources should be examined, as should
the possibility of dysfunctional vacuum
pumps or gages.
Gases and vapor can be released inside
the system from the chamber walls and
other materials inside the system
(outgassing) or from small volumes with
very low conductance paths for pumping
(virtual leaks). Outgassing results from the
evaporation of materials in the vacuum
system (e.g., organic materials, ice on the
exposed surfaces of cold traps and
elsewhere, oil or grease etc.) as well as
from permeation through the walls of the
vessel and desorption of gas and vapor
from interior surfaces. Outgassing is best
controlled by careful attention to the
properties of materials permitted in the
system, cleanliness in construction and
use of the system and the use of bakeout
and cold trap techniques. Virtual leaks
commonly arise from double welds,
double gasket design, blind stud holes
that are not vented etc. and can be
avoided by proper design and fabrication.
The various considerations and
techniques used to minimize outgassing
and virtual leaks are described earlier in
this chapter.
Analysis of Vacuum System
Pressure Transients during
Pumpdown and without
Pumping
Some degree of outgassing will be present
in any vacuum system and will constitute
a larger proportion of the gas pumped out
as the vacuum decreases. An indication of
the amount of condensable vapor present
can be obtained from vacuum gage
readings made with and without a cold
trap. A marked reduction in pressure
when the cold trap is filled indicates the
presence of condensable vapors arising
from outgassing surfaces and virtual leaks.
FIGURE 33. Pressure versus time curve of vacuum system
pumpdown and subsequent measurement of rise rate.
F
A
Leaks
Pumpdown
curve
Pressure
Leaks, Outgassing and
Trapped Gas (Virtual
Leaks) in Operating
Vacuum Systems
Measurement of the rate of pressure rise
can be used to verify the presence of leaks
and can also provide an estimate of their
size if the volume of the system is known.
Figure 33 shows a typical pressure time
curve for a vacuum system with a liquid
nitrogen cold trap. The curve shows the
pressure variations during the pumpdown
cycle and during a rate-of-rise
measurement. The characteristic
exponential decrease in pressure occurs
from A to B, during pumpdown. The
pressure levels off as the system
approaches equilibrium between pumping
speed and the gas load from leaks and
outgassing. At B the liquid nitrogen trap is
filled and the pressure falls rapidly as
condensable vapors are captured by the
trap. Again an equilibrium pressure is
reached, limited by noncondensable gas
from leaks. At point D the vacuum
chamber is valved off from the pumps
and cold trap and the pressure begins to
rise. The rate of pressure rise will decrease
in the region from D to E as the
contribution from outgassing becomes
negligible in comparison with any leaks
present. Finally, the pressure-time curve
becomes nearly a straight line in region
E-F. If slope dP/dt approximates Q L/V,
where Q L is the leakage rate and V is the
volume of the vessel.
E
Liquid nitrogen
applied
Outgassing
and leaks
B
Valve closed
Vapors
(mostly)
C
D
Time
Legend
A = Pressure before pumpdown
B = Liquid nitrogen trap is filled
C = Trap captures condensable vapors
D = Vacuum chamber is valved off
E = Pressure rise curve is no longer influenced by outgassing
F = Final reading
Leak Testing of Vacuum Systems
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
255
LT.06 LAYOUT 11/8/04 2:18 PM Page 256
Sensitivities of Leak Tests
for Operating Vacuum
Systems
The choice of leak testing technique for
use on operating vacuum systems depends
on such factors as (1) magnitude of
leakage, (2) pressure within vacuum
chamber during leakage detection and/or
measurement, (3) pressure external to
vacuum chamber during leakage detection
and/or measurement, (4) capacity of
vacuum pumps at operating pressure with
leakage occurring, (5) tracer gas type and
ease of detection (if tracer is other than
air), (6) internal volume of vacuum
system, (7) virtual leakage and effects of
outgassing and (8) sensitivity of vacuum
gage or tracer leak detector used in leak
testing.
Table 4 lists the pressure ranges and
leakage rate sensitivities of various
techniques of leak testing of operating
vacuum systems. Of course, when vacuum
systems are not operating and can be
pressurized or when components of
vacuum systems can be removed and
tested separately for leaks, the many other
leak testing techniques described in this
volume may be applicable.
Auditory Aids to Detection
of Large Leaks in
Operating Vacuum
Systems
The first indication of the existence of a
large leak in a continuously pumped
vacuum system is usually an audible
one—the distinctive sound of a
mechanical pump that is pumping large
quantities of air long after the system
should be in the initial vacuum range (see
curve AB in Fig. 33). Gross leaks
correspond to openings with diameters of
about 10 µm (4 × 10–4 in.) and larger.
Hence, the hissing of air through large
leaks can sometimes be heard and used to
locate them. An improvised stethoscope
or listening tube improves both the
sensitivity of the technique and the
ability to pinpoint the location of the
leak. Advanced ultrasonic leak detectors
can also be used to locate large leaks.
Sensitivity may also be improved (and the
pump spared) if pressure testing is used
instead of vacuum testing.
TABLE 4. Sensitivities of some techniques of leak testing in vacuum systems.
Smallest Detectable
Pressure
Range
Leakage
__________________________________
_______________________________________
Technique
Hissing of air
Wavering flame
Halide torch
kPa
10 to 200 kPa
100 to 400 kPa
>100 kPa
(torr)
(100 to 2000)
(1000 to 4000)
(1000)
Bubble techniques
air and water immersion
water and alcohol immersion
air and soap film
0.01 to 400 kPa
0.01 to 400 kPa
0.01 to 400 kPa
(1 to 4000)
(1 to 4000)
(1 to 4000)
Spark coil or discharge tube
0.1 to 100 Pa
(0.001 to 1.0)
Pa·m3·s–1
(std cm3·s–1)
3×
4 × 10–3
1 × 10–5
(3 ×
(4 × 10–2)
(1 × 10–4)
1.5 × 10–5
5 × 10–8
5 × 10–6
(1.5 × 10–4)
(5.0 × 10–7)
(5.0 × 10–5)
10–3
~0.001
10–2)
(~1 × 10–2)
Pirani and thermocouple gages <1 × 101 Pa
(0.1)
1 × 10–6 to 1 × 10–7 (1 × 10–5 to 1 × 10–6)
Halogen detector
Ionization gage
<10 Pa
<0.07 Pa
(0.1)
(0.0008)
1 × 10–7
(1 × 10–6)
dependent on pressure
<0.01 Pa
(0.0001)
Mass spectrometer leak detector
direct flow
<0.01 Pa
counterflow
40 Pa
residual gas analyzer
(0.0001)
(0.3)
Ion pump leak detector
256
Leak Testing
dependent on pressure
5 × 1012
1 × 1011
10–10 to 10–11
Remarks
quiet room
draft-free room
used with
refrigerant-12
good ventilation
good light; ≥ 5 min
observation
leakage dependent
on voltage; glass system;
residual
gases cause
confusion
used with acetone,
hydrogen methanol
used with hydrogen,
helium, oxygen,
butane
used with argon,
oxygen, helium
(5 × 1011)
used with helium
(1 × 1010)
used with helium
(1 × 10–9 to 1 × 10–10) used with any gas
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
Pressure Gage Leakage
Tests of Small Vacuum
Systems in Operation
A simple technique can be used for
preliminary leakage testing of small high
vacuum systems in operation. This
technique makes use of the vacuum gage
that already exists in most vacuum
systems. The most common gages are of
the thermal conductivity type for
pressures as low as about 0.1 Pa (1 mtorr)
and some variation of the ionization gage
for pressures below the 0.1 Pa (1 mtorr)
range. Both gage types can be used for
leakage detection, but the ionization gage
is preferable because of its faster reaction
time. However, if a very large leak makes
it impossible for the pump to reach the
working range of the ionization gage, the
thermocouple gage may be used in
essentially the same way but at a slower
pace.
High Pressure Air Jet Tracer
Technique for Locating Leaks in
Operating Vacuum Systems
A simple leak locating tracer technique
involves blowing a jet of high-pressure air
onto the outside of the vacuum chamber
wall. This raises the air pressure
differential across a small area of the
chamber wall. If a leak is within this area
it will now conduct more air into the
chamber. The higher leakage rate can
immediately be detected on the vacuum
gage as a slight increase in chamber
pressure.
In practice, a sharp air jet from a small
nozzle is moved over all suspected areas;
the common shop air supply system will
do very well. The scanning can be rapid,
because reaction and recovery times are of
only a few seconds duration. This
technique is most useful for quickly
testing for leaks in a weld or an O-ring
sealed flange.
Vacuum Hose Technique for
Locating Leaks in Small Operating
Vacuum Systems
Another simple technique of locating
leaks in operating vacuum systems is
based on the same idea, to change the
pressure differential across the leak and to
observe the change in leakage rate with
help of the gage. This time, however, the
pressure on the air side of the leak is
reduced rather than increased. For this
procedure, a source of vacuum is required.
The vacuum line available in many
laboratories, a small vane pump or even a
water injection pump are all adequate. A
hose of appropriate diameter is connected
to the vacuum pump; its other end is
then simply pressed against the chamber
wall to create a small area of reduced
pressure. If the leak is within this area, an
almost immediate improvement in the
chamber vacuum will result.
The vacuum hose technique works best
on flat, smooth wall sections. Its special
merit, besides being very fast, is that the
area under investigation is sharply limited
and very well defined. In cases where
there are several potential leaks in a small
area, it has proven to be superior to any
tracer gas technique. The vacuum hose
can be applied to one small zone after
another until the leak is positively
localized, whereas it is difficult to confine
any tracer gas to equally small zones
without diffusing some into adjacent
areas.
Helium Mass Spectrometry
The helium mass spectrometer leak
detector (usually referred to simply as a
helium leak detector) is adjusted to
respond only to helium gas
(atomic mass = 4). Although several types
of mass spectrometer are used in these
devices, by far the most common is the
simple magnetic analyzer.
By choosing the suitable magnetic field
strength and acceleration voltage, the
mass spectrometer can be tuned to any
mass of gaseous particle. Hence, any gas
could be used as a tracer gas for leak
detection. Helium has often been chosen
for the following reasons. It is present in
the atmosphere at a concentration of
about 5 µL·L–1. Thus, air leaks cause very
little helium background in the detector.
Helium is inert and readily available in
most countries. Because it is the lightest
gas except hydrogen, helium’s diffusion
and molecular flow rates are the highest
available with a nonhazardous gas. These
properties are highly desirable in a tracer
gas.
Helium Tracer Gas for Large Leaks
in Vacuum Systems
The helium mass spectrometer leak
detector can sometimes be used to find
even large leaks, although its main use is
in finding small and very small leaks.
Because the pressure in the conventional
helium mass spectrometer leak detector
cannot exceed 10 mPa (0.1 mtorr), the
leaking vacuum system is pumped at the
greatest attainable pumping speed and the
opening to the leak detector is then
throttled until the operating pressure is
achieved.
It is particularly important that the
helium probing procedure be observed
when testing for large leaks. Otherwise,
Leak Testing of Vacuum Systems
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
257
the detector can easily become saturated
long before the leak is reached. If a
counterflow leak detector is being used,
testing pressures maybe as high as 40 Pa
(0.3 torr) can be tolerated without the
need for throttling the leak detector,
without loss of sensitivity. Some models
of leak detectors have built-in capabilities
of testing at pressures as high as 400 Pa
(3.0 torr). The pressure testing technique
on large leaks is virtually impossible
because it saturates the mass spectrometer
detector chamber with helium tracer gas.
Leak Testing of Vacuum
Systems of from 100 to
0.1 Pa (100 to 1 mtorr)
Most of the above techniques for
detecting large leaks have sufficient
sensitivity to be useful with leaks that
limit the pressure to the vacuum range of
100 to 0.1 Pa (100 to 1 mtorr) with the
pumping speeds commonly used in this
range (S ≥ ~1 L·s–1 or ~2 ft3·min–1).
However, when vacuum system pressures
lower than 100 Pa (1 torr) can be
obtained, several additional vacuum leak
testing techniques avoid the
inconvenience of pressure testing and can
be used on systems that cannot be
pressure tested.
Tesla coils and high voltage discharge
devices, which were among the earliest
leak detection tools used on vacuum
systems, provide a rather qualitative
indication of the pressure and type of gas
in the system. They can be used only on
glass systems or in glass walled sections of
metal systems. For example, they can be
used along the glass tube leading to an
ion gage only if the ion gage is turned off.
Commercial spark coils (Tesla coils) for
vacuum testing produce a high frequency
potential of several thousand volts at a
pointed electrode. When the tip of this
electrode is held near (about 1 cm from) a
glass system whose pressure is in the
vacuum range of 100 to 0.1 Pa (1.0 to
0.001 torr), a gaseous electrical discharge
is produced in the vicinity of the
electrode. The color and appearance of
this gaseous discharge depend on the
composition of the residual gas in the
system and on its pressure.
Sensitivity and Limitations of
Spark Coil Leak Location
The white spark technique of high voltage
discharge leak location is qualitative, but
will probably detect leakage as small as
10–5 Pa·m3·s–1 (10–4 std cm3·s–1). The size
of the smallest detectable leak depends on
leak geometry. The leak testing technique
consists of evacuating the system to a
258
Leak Testing
pressure between 1 kPa and 1 Pa (10 and
0.01 torr) and scanning over the
suspected areas with a probe connected to
a high voltage induction coil.
The white spark technique is only
applicable where no metal exists because
the spark from such a coil will ground
through metal parts. If the spark tip is
brought closer than several centimeters
from metal parts, the spark will jump to
the metal. Thus, spark coils cannot be
used on all-metal systems. However, they
can be quite useful on all-glass systems or
even on metal systems containing glass
parts. On continual exposure, the high
voltage spark may puncture thin glass
walls. Therefore, the probe should be
moved slowly rather than held in one
place. In the same manner, a high voltage
spark might score the barrel of a
fluorocarbon resin stopcock and rupture
plastic or rubber gaskets.
Location of Vacuum System Leaks
by Glow Discharge Color
The color differentiation technique of
high voltage discharge leak testing is
primarily a technique for leak location
and is applicable to evacuated systems. It
is always used in the tracer probe mode.
The color differential technique involves
observing changes in color of high voltage
glow discharges within the evacuated
space produced by probe gases or vapors
entering the leak. A spark coil can be used
to excite a visible glow discharge if the
pressure in the system is within the range
of 1 Pa to 1 kPa (0.01 to 10 torr). A tracer
gas such as carbon dioxide or a volatile
liquid such as benzene, acetone or methyl
alcohol is applied to the exposed outer
surface of the vacuum system under test.
When the tracer gas or vapor enters the
system through a leak, the color of the
discharge changes from the reddish purple
of air to a color characteristic of the tracer
material. For liquids such as benzene,
acetone or alcohol, the color of the glow
discharge would be grayish blue. Carbon
dioxide gives a bluish green glow to the
electrical discharge.
During glow discharge leak testing of
vacuum systems, the spark coil tip is kept
on one glass section of the system under
test. Preferably this section will be
between the diffusion pump and the
forepump to have a pressure sufficient to
maintain a glow discharge. The nature of
the glow discharge will depend on the
pressure and on the gases in the system.
The glow discharge color is characteristic
of the gases present. For air, this color is
reddish or purplish. The exact color (as for
other gases) depends to some extent on
the glass used in the system. Soda glass
will show a yellow-green fluorescence
whereas lead glass shows a blue
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
fluorescence. The probing fluid used can
be a gas or a liquid. Some tracer materials
that are commonly used are illuminating
gas, ether and carbon dioxide. With the
first two materials the discharge takes on
a grayish blue appearance. This is similar
to the characteristic color or carbon
dioxide (see Table 5) but, possibly because
of fluorescence of the glass, the color is
often reported as bluish green.
Leak Location by Isolation
in Operating Vacuum
Systems
The principle of fault isolation applies
particularly strongly to leak detection in
operating vacuum systems. Because leak
hunting is usually a tedious and time
consuming job at best, any steps taken to
isolate the leak to a particular part of the
system can shorten the process
considerably. Often various parts of the
system can be valved off and pressure
gages used to indicate when the leak has
been isolated. A system that has a history
of achieving adequately low pressure may
leak after being opened. In this case, the
obvious initial candidates for leak testing
are the gaskets on any flanges removed
and possible the valves used to vent or
seal off the system. For many systems
there is a high probability that the leak
will be found in these mechanical seal
areas rather than elsewhere, but in some
cases, such as when temperature cycling
of the system is involved, the new leaks
may be far removed from the openings
TABLE 5. Discharge colors in gases and vapors at low
pressures.
Gas
Air
Nitrogen
Oxygen
Hydrogen
Helium
Argon
Neon
Krypton
Xenon
Carbon
monoxide
Carbon
dioxide
Methane
Ammonia
Chlorine
Bromine
Iodine
Lithium
Sodium
Potassium
Mercury
Negative Glow
Positive Column
blue
reddish
blue
yellow or red gold
yellowish white
lemon
bluish pink or bright blue pink or rose
pale green
violet-red
bluish
deep red or violet
red-orange
red-orange or blood red
green
no distinctive color
bluish white
greenish white
white
blue
reddish violet
yellow-green
greenish
yellowish green
orange-yellow
bright red
yellowish green (whitish)
green
green or goldish white
white
light green
reddish
peach blossom colored
used. If the leak can be positively isolated
to a given area, it should be.
Sealing Technique for
Determining Leak Location
The sealing technique involves gradually
covering outside parts of a system being
evacuated with some material that will
seal the leak. Once the leak has been
covered, the pressure will drop. In this
way, leaks can be located and permanent
repairs made. The procedure is to paint,
brush or spray the sealing substance over
various parts of the system until a change
in pressure is noted. Either a thermal
conductivity or an ionization gage may be
used, the choice being dictated by the
pressure.
The sealing substance may temporarily
or permanently seal the leaks. some
semipermanent sealants are insulator
lacquers, shellac in alcohol, epoxy and
vacuum cements that are liquid at room
temperature such as cellulose acetate.
Some temporary sealants are water,
acetone and alcohol.
Two effects result from a liquid sealant.
First, after the initial closing of the leak,
the pressure will drop. Second, as the
vapor enters the system, the gage will
show a change in pressure, which will
depend on the nature of the vapor and on
the type of gage. The vapors from solvents
such as water, acetone or alcohol are
readily condensable. Consequently, all
gages used with a cold trap will show a
pressure change when a leak is covered by
a liquid. The particular liquid used (no
cold trap) will determine whether the
gage shows an increase or decrease in
pressure. Alcohol, acetone and ether —
commonly used probe liquids — all show
an initial increased pressure reading with
an ionization gage or thermal
conductivity gage but may then change to
a decrease in pressure due to the
temporary plugging of the leak.
Effect of Sealant Material with
Very Small Leaks
For very small leaks, a permanent
sealing material works satisfactorily. The
temporary sealing substances are quite
effective for all sizes of leaks except the
very smallest. If a very small leak is sealed
with a temporary sealant, it will open
again at some inopportune time;
therefore, this technique is not
recommended if the small leaks have to
be located and permanently repaired.
yellow
green
greenish blue or greenish
Leak Testing of Vacuum Systems
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
259
Effect of Sealant Material with
Large Leaks
In repairing large leaks, the sealant
material is drawn into the vacuum system
and a seal cannot be obtained. Although
the permanent sealing substances give
fairly satisfactory results with leaks in
metal plates, in soldered, brazed and
welded joints and in glass systems, they
are not as satisfactory as a final repair
obtained by reworking the material of the
vacuum system by soldering or welding.
The permanent sealing substances make
further reworking of the glass or metal
very difficult.
Temporary Sealants to Locate
Large Leaks in Vacuum Systems
Despite its drawbacks, the traditional
technique of sealing suspected leak areas
can sometimes succeed where other
techniques fail It involves the application
of a low vapor pressure sealant (usually
vacuum putty or duct seal) to the
suspected leak. The process is time
consuming. It can cause difficulty in
making a permanent leak repair unless the
sealant is all removed with solvent before
repairs are made. In no event should
vacuum putty or other sealants be relied
on for a permanent seal.
A leak can in effect be sealed by
applying a forevacuum to the region
external to the suspected leak. For
example, a flange joint can be sealed with
tape except for a gap at one point. A
vacuum hose can then be pressed against
this gap to evacuate the volume around
the flange gasket. Although obviously
limited in scope, this overvacuum
technique can be useful in leak isolation.
Repairs of Large Leaks in
Operating Vacuum
Systems
If any general advice can be given about
the repair of leaks, design can help
considerably in reducing exposed areas.
Because the outgassing rate of
elastomers increases as the temperature is
raised, the ultimate pressure can be
reached more rapidly if the elastomer can
be heated. However, all elastomers are
damaged when heated too much. Also,
the compression set increases more
rapidly with temperature.
Because of these properties, elastomeric
gaskets are not normally used in ultrahigh
vacuum systems. Such systems are baked
at temperatures well above the damage
point of insulator lacquers, sealing waxes,
fast setting adhesives, epoxy coatings,
vinyl plastic coatings, solder (and
260
Leak Testing
glycerine at liquid helium temperatures).
In short, temporary leak seals are made
with almost anything handy. Some
techniques, e.g., epoxy, come close to
being permanent repairs, but most
temporary seals can be expected to give
trouble at some further time. They can be
a constant source of worry if not properly
repaired when it first becomes possible to
make a permanent seal.
The simplest leak to repair properly is a
leaking flange gasket that can be sealed
either by tightening the flange bolts a
little more or by replacing the gasket.
Most other leaks require reworking of the
part. Leaking welds should be ground
down to a smooth, clean surface before
rewelding to help prevent the formation
of a virtual leak under the new weld. In
all cases, all vestiges of any temporary
sealants used must be removed before
starting a repair.
Sensitivity of Glow
Discharge Color Leak
Testing
The color differentiation technique will
detect a gas pressure change of about 1 Pa
(10 mtorr). The sensitivity of the
technique is dependent on the pumping
speed of the vacuum system as measured
in the glow discharge area.
Limitations of Glow
Discharge Color Leak
Testing Technique
Part of the vacuum envelope of the
system under test has to be transparent so
that the change in color of the discharge
can be seen when leaks exist. Because the
procedure depends on detecting total
tracer gas pressure buildup, the time that
the test object has to be left standing
before testing increases with an increase
in desired leakage sensitivity. Any gas or
liquid whose glow discharge color is
different from the background discharge
color may be used as a tracer. However,
gasoline, benzene, pyridine and solutions
containing nitrogen compounds should
not be used as tracers because they adhere
to glass.
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
PART 7. Leak Testing of Vacuum Systems by
Vacuum Gage Response Technique
Principles of Leak Testing
by Vacuum Gage Response
to Tracer Gases
The procedure of leak testing by vacuum
gage response is based on the principle
that most vacuum gages have a pressure
response dependent on gas composition.
If the composition of gas in a system
changes, the reading on the gage reflect
this change. Leak location therefore
consists of spraying a tracer gas on the
suspected leak and observing any response
by the vacuum gage to the tracer gas that
enters the system through the leak.
Most stainless steels used in vacuum
work are called 18-8 stainless steels
because they contain about 18 percent
chromium and 8 percent nickel. These
steels are nonmagnetic and the melting
points of austenitic stainless steels are
over 1400 °C (2550 °F). Surfaces of
stainless steels remain smooth because
oxides and hydroxides do not occur as in
other types of metals. This means that the
effective surface area is less and vapors are
adsorbed in smaller quantities. This leads
to much easier degassing and quicker
pumpdown.
The vacuum gage leak test depends on
maintaining a constant gas pressure in the
system. If the system pressure varies for
reasons unrelated to testing, leak location
using pressure gages is impossible. The
sensitivity of vacuum gage leak testing is
relatively low (10–5 Pa·m3·s–1 or 10–4 std
cm3·s–1). The necessary instruments
cannot be used in a contaminated
atmosphere because they will respond to
other gases present in the air. Therefore,
these instruments are not widely used
where welding (inert gases), cleaning
(solvent fumes), brazing (combustion
products) or painting (paint solvents)
operations are performed.
Rubber and grease should be
minimized, particularly in the connection
link to the leak test gage being used as the
detector, because they tend to absorb
tracer gas (helium, halogens etc.) in the
early phases of leak testing and outgas
them later when high sensitivity is
needed.
Procedures for Locating Leaks by
Vacuum Gage Tests
In the evacuation mode, the system under
test is evacuated and the suspected leak is
sprayed with tracer gas (see Fig. 34).
Pressure gage response to the tracer gas
indicates that a leak has been located. The
procedure is to expose small areas of the
external pressure boundary surfaces of an
evacuated system to a tracer gas. If a leak
is present, this gas enters the evacuated
system and displaces or mixes with any
residual gas in the neighborhood of the
gage. There are several variations of this
procedure, depending on the vacuum
gage used and the technique of increasing
specificity, but the various techniques
have a number of feature in common.
Application of Vacuum
Gage Leak Testing
The vacuum gage leak testing procedure is
extremely popular for leak location on
vacuum systems because a pressure gage is
usually built into the system. The only
other requirement for the test is tracer
gas. This procedure was once widely used
for leak testing of components, but with
the advent of more specific and more
FIGURE 34. Idealized system for vacuum
gage response testing.
Tracer
probe gas
Leak Q
P
System
being
tested
Gage
Conductance C
Volume V
Diffusion pump:
speed s
Leak Testing of Vacuum Systems
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
261
sensitive leak detectors, it has fallen into
disuse.
It is possible to use the vacuum gage
response leak testing procedure for
approximating leakage measurement on
vacuum system. This is done by
stabilizing the system, hooding it and
introducing tracer gas into the hood.
However, the response is not generally
quantitative and is too nonspecific to be
of much value. It is always questionable
whether the pressure age response is due
to increased concentration of the tracer
gas or to some other factor.
Sensitivity of Vacuum Gage Leak
Testing
The sensitivity of vacuum gage leak
testing is dependent on the sensitivity of
the absolute pressure gages being used
and on the pumping system on which
they are mounted. The leakage sensitivity
is ordinarily in the range of 10–5 to
10–7 Pa·m3·s–1 (10–4 to 10–6 std cm3·s–1).
This can be increased by modifications
that increase specificity of the gage
response to the tracer gas.
In the tracer probe leak testing
technique, the size of the leak that can be
detected by a vacuum gage is dependent
on the pumping speed of the system. As a
first approximation, this procedure can
detect a pressure change of one fiftieth of
the pressure in the system. Smaller leaks,
i.e., leaks that do not contribute more to
system pressure or composition, will not
be detected by this procedure.
Characteristics of Typical Vacuum
Gages Used in Leak Testing
Many gages such as the Pirani and
thermocouple gages use the thermal
conductivity principle to measure
pressure. These gages usually have a leak
checking position on their meter scale. In
this position, the pointer is in the center
of the meter scale and operates at high
sensitivity. Any movement of the pointer
indicates a leak. Some instruments
amplify the change of pressure indication
of gages, which simplifies leak location
procedures. Ionization gages are
specifically modified for leak testing of
evacuated systems.
Advantages of Leak Testing with
Vacuum Gages
The major advantage of leak testing with
vacuum gages on existing vacuum systems
is that no additional leak testing
equipment is necessary. Leak location may
be performed using gages already on the
system. The procedure is inexpensive and
does not require highly trained test
personnel. In the pressurizing mode, leak
262
Leak Testing
testing by thermal conductivity gage
response is also an inexpensive technique
of leak location. The equipment is
portable and may be used on a variety of
gases in the system.
Maximum Sensitivity of
Leak Testing by Vacuum
Gage Response
Maximum sensitivity will be obtained
when the test includes (1) complete
coverage of the leak by the tracer gas;
(2) high sensitivity of the gage to the
tracer gas; (3) low value of viscosity of the
tracer gas; (4) a small effective pumping
speed for the tracer gas; and (5) tracer gas
with a high molecular weight.
Effect of Selection of
Vacuum Pump
It is possible to use a small pump to
evacuate the system being tested, but
pressure fluctuation will be created. The
pumping speed is more effectively
reduced by using a large pump and a
small conductance connection to the
system. In practice, a turbomolecular or
diffusion pump is preferable to a
mechanical pump, because these pumps
produce less pressure fluctuation. Of
course, on a system with built-in pumps
the pumping speed can not be altered for
leak location, so the sensitivity is fixed by
system design.
Effect of Molecular Flow
In-Leakage on Vacuum
Gage Response
For gage response for large leaks, it can be
assumed that flow through the leak is
laminar. In small leaks (10–7 Pa·m3·s–1 or
10–6 std cm3·s–1), the flow will be
molecular. In molecular flow, the leakage
is inversely proportional to the square
root of the molecular weight of the
leaking gas. The same relationship applies
to the conductance that determines the
pumping speed of tubulation (see Eq. 27).
If the leakage into the system is molecular
and the pumping speed is determined by
the tubulation leading to the pump, the
pressure in the system is independent of
the property of the leaking gas. The gage
response is then dependent only on the
relative sensitivity of the gage to the
tracer gas as compared to air.
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
LT.06 LAYOUT 11/8/04 2:18 PM Page 263
techniques and (3) ionization efficiency
techniques. These in turn have various
subdivisions.
Tracer Gas Pressure
Sensitivity Factor for
Vacuum Gages
Because there are a variety of factors
involved in choosing a combination of
proper tracer gas and vacuum gage, it is
often easier to determine the sensitivity
factor experimentally:
Pressure caused by
(30)
φ
tracer gas on the leak
Pressure on system
=
with air on leak
The experimental values of this tracer gas
sensitivity factor are listed in Table 6.
The minimum detectable leakage can be
determined from tracer gas sensitivity
factor and leak testing conditions:
(31) Q min
∆ P2 S a
φ
=
where ∆P2 is smallest measurable air
pressure variation, Qmin is smallest
measurable leakage, Sa is pumping speed
for air at the gage and φ is ratio defined by
Eq. 30.
It is apparent from the above
discussion that the minimum measurable
leakage will be within a decade of the
minimum measurable pressure change,
multiplied by the pumping speed at the
pressure measurement site. In designing
this type of leakage measurement, the
response time of the system must also be
taken into account. The response time
constant Tc of the system is the time for
the leak indication to fall to 1/ or
36.4 percent to its maximum value.
(32) Tc
=
Factors Affecting
Sensitivity and Response
Time of Vacuum Gage Leak
Testing
The pumping speed Sa used in Eq. 31 is
the pumping speed at the site of the gage.
Thus, the location of the gage affects the
sensitivity. If the gage is connected by
way of a restriction, it will be difficult to
detect small leaks anywhere except near
the gage itself.
The response time depends primarily
on the volume V of the system and on
the effective pumping speeds at the test
chamber, i.e., on the speeds S for air and
KS for the tracer gas. The pumping speed
of a turbomolecular or diffusion pump
varies inversely as the square root of the
molecular mass. The vacuum gage
response will depend on the ratio of the
leak detector response for air to its
response for the tracer gas. The gage
response will also depend on the ratio of
the leakage rate for tracer gas to the
leakage rate for air.
V
KS
where V is the volume of the evacuated
system, K is the ratio of effective pumping
speed for tracer gas to pumping speed for
air and S is pumping speed.
The testing techniques can be divided
into three categories: (1) sealing
techniques, (2) thermal conductivity
TABLE 6. Tracer gas sensitivity factor.
Tracer Gas
Butane
Diethyl ether
Carbon dioxide
Carbon
tetrachloride
Benzene
Hydrogen
Coal gas
Hot Cathode
Ionization Gage
Pirani Gage
10.0
5.0
1.0
1.0
0.7
0.3
1.0
0.3
0.4
0.25
0.05
0.1
0.4
0.25
Leak Testing of Vacuum Systems
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
263
PART 8. Leak Testing of Systems by Thermal
Conductivity Techniques
Thermal Conductivity
Technique for Leak Testing
of Vacuum Systems
The thermal conductivity leak testing
technique can be used with either the
pressurized system (detector probe)
technique or the evacuated system (tracer
probe) technique. In the evacuated system
mode of leak testing, gages normally
found on the system are used. In the
pressurized system mode, special leak
detectors are necessary.
Tracer Probe Technique of
Thermal Conductivity Leak
Testing
The tracer gas detector for the tracer probe
technique all evolved from thermal
conductivity gages present on vacuum
systems. Either thermocouple or Pirani
gages normally mounted on the vacuum
system are used for thermal conductivity
leak testing by the tracer probe technique.
Because these gages best respond to a
pressure between 100 Pa and 10 mPa
(1 torr and 0.1 mtorr), they are used on
systems with low pumping speed.
Alternatively, these gages can be placed
between the turbomolecular or diffusion
pump and the fore pump on a vacuum
system. The thermal conductivity
technique is very old, yet it is continually
used in leak location on vacuum systems.
New tracer fluids are used to enhance the
technique and modifications are made on
the pumping equipment to increase the
leakage sensitivity.
Because the response of a thermal
conductivity gage depends on the mass of
the gas molecules, these gages can be used
with a tracer gas to find leaks. When a
leak is covered with a light gas such as
helium, the gage will read higher than for
an air leak. Conversely, a heavy gas such
as argon will cause the gage reading to
decrease. Volatile liquids such as acetone
or alcohol can also be used but the
response will depend on whether the
vapors enter the leak or the liquid freezes
in the leak, temporarily sealing it. One
must keep in mind that (because of the
fairly long response time of thermal
gages) the leak may have been covered
264
Leak Testing
some time before the gage gives any
indication. Hence, the leak may have to
be located by successive approximations
— a characteristic of most leak detection
techniques. Because most vacuum systems
will have either a thermocouple or Pirani
gage to monitor fore pressure, these gages
in the pressure range from 0.1 to 30 Pa
(1 to 300 mtorr) are both simple and
convenient.
Thermal Conductivity Leak
Testing with Hydrogen
Tracer Gas and Charcoal
Trap
For example, if probing with hydrogen
gas, an increase of tracer gas partial
pressure may be obtained by reducing the
turbomolecular pump speed with an inbleed or reducing the diffusion pump
speed by reducing the heater voltage. This
decrease of hydrogen gas pumping speed
is obtained without materially reducing
the pumping speed for other gases.
Modifications of this simple leak location
technique are similar to those described
later in this chapter for ionization gages.
For example, in a Pirani leak detector
using hydrogen gas, the gage is isolated
from the system by a cooled charcoal trap.
With this device it is possible to locate
leaks as small as 10–7 Pa·m3·s–1
(10–6 std·cm3·s–1).
Thermal Conductivity Leak
Testing with Butane Tracer
Gas
A differential leak detector for butane
tracer gas uses two vacuum gages in a
Wheatstone bridge circuit. One of the
gages is in series with a charcoal trap. This
arrangement has stability because any
random pressure changes will be detected
by both gages while the butane tracer gas
will be absorbed by the charcoal. In this
technique, the charcoal does not have to
be heated during detection. The
sensitivity of this differential system is
reported to be 10–7 Pa·m3·s–1
(10–6 std cm3·s–1). Some thermal
conductivity leak detectors are specifically
designed for the detector probe technique.
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
Thermal Conductivities of
Different Tracer Gases
Effect of Detector Probe
Pumping Speed
In principle, any tracer gas having a
thermal conductivity different from that
of air could be used with thermal
conductivity leak detectors. The leakage
sensitivity depends on relative differences
of the thermal conductivities of the gases
that are compared in Table 7. It is
apparent that both hydrogen and helium
show large relative differences and are
therefore the most sensitive tracer gases
with this technique. For special
applications, it is sometimes desirable to
use one of the other tracer gases. Table 7
gives some indication of results expected.
It is clear that either gases with a thermal
conductivity greater than air (such as
helium, methane etc.) or those with
thermal conductivities less than air (such
as halogenated hydrocarbons, argon,
carbon dioxide etc.) would be suitable.
The tracer gas emerging from leaks is drawn
into sampling probes by the action of a
small pump. The pump can be run at two
speeds: a maximum speed for fast response
and a slower speed to give an increased
detection sensitivity at some sacrifice in
response time. To obtain a good response,
the thermal conductivity sensing elements
must be small enough to fit in chambers of
small volume. Because it is intended to
detect changes in gas concentration rather
than rates of flow, the gas should be made
to flow past the entrance of the element
chambers rather than through them.
Thermal Conductivity Leak
Detector with Hot Wire
Bridge Sensor
The thermal conductivity leak detector of
Fig. 35 is based on a hot-wire bridge in
TABLE 7. Thermal conductivities of tracer gases for a temperature 20 °C (70 °F) in units
of W·m–1· K–1 (BTU·h–1·ft–2·°F–1·ft).
Chemical
Formula
Gas
Air
Acetylene
Ammonia
Argon
Benzene
Butane
Carbon dioxide
Carbon disulfide
Carbon monoxide
Ethane
Ethylene
Halogenated hydrocarbon
Halogenated hydrocarbon
Halogenated hydrocarbon
Halogenated hydrocarbon
Halogenated hydrocarbon
Halogenated hydrocarbon
Halogenated hydrocarbon
Helium
Hydrogen
Hydrogen sulfide
Krypton
Methane
Neon
Nitric oxide
Nitrogen
Nitrous oxide
Oxygen
Propane
Sulfur dioxide
Water vapor
Xenon
F-11
F-12
F-21
F-22
F-113
F-114
F-132
Molecular Mass
(atomic mass units)
Thermal Conductivitya
_______________________
BTU·h–1
________
W·m–1· K–1
ft2·°F·ft–1
mixture
C2H2
NH3
A
C6H6
C4H10
CO2
CS2
CO
C2H6
C2H4
CCl3F
CCl2F2
CHCl2F
CHClF2
CClF-CClF2
CClF2-CClF2
29.9
26.0
17.0
39.9
78.0
58.0
44.0
76.0
28.0
30.0
28.0
137.4
120.9
102.9
86.5
187.4
170.9
0.025 57
0.019 51
0.023 06
0.017 58
0.009 31
0.014 22
0.015 10
0.007 10
0.023 53
0.019 06
0.017 73
0.008 13
0.009 58
0.011 42
0.007 58
0.010 88
0.151 20
0.014 78
0.011 28
0.013 33
0.010 16
0.005 38
0.008 22
0.008 73
0.004 10
0.013 60
0.011 02
0.010 25
0.004 70
0.005 42
0.005 54
0.006 60
0.004 38
0.006 29
He
H2
H2S
Kr
CH4
Ne
NO
N2
N2O
O2
C3H8
SO2
H2O
Xe
4.0
2.0
34.0
83.8
16.0
20.2
30.0
28.0
44.0
32.0
44.0
64.0
18.0
131.3
0.186 32
0.013 32
0.009 34
0.032 39
0.046 02
0.020 41
0.025 29
0.016 00
0.025 78
0.016 00
0.025 78
0.016 00
0.018 81
0.051 90
0.087 40
0.107 70
0.007 70
0.005 40
0.018 72
0.026 60
0.011 80
0.014 62
0.009 25
0.014 90
0.009 25
0.005 14
0.010 87
0.030 00
a. Thermal conductivity values for a temperature of 20 °C (70 °F) in units of W·m–1·K (BTU·h–1·ft–2·°F–1·ft).
Leak Testing of Vacuum Systems
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
265
which two resistance elements form two
arms of the bridge network. One element
is exposed to air containing tracer gas,
while the other is exposed only to air and
serves as a reference to compensate for
changes in ambient conditions. As shown
in Fig. 35, the sensing elements are
mounted in a metal block inside a handheld probe unit. Gas samples are drawn
up through a narrow-bore tube. The
sensing elements consist of coils of thin
tungsten wire mounted on glass-metal
seals in a compact assembly, into which
the pump connects. The sensing probe is
also fitted with a small meter to repeat the
leak indication of the amplifier unit.
Operators find this assembly to be
convenient, particularly when testing
awkwardly shaped equipment.
The electronic circuitry can be
transistorized and thereby made compact
enough for the unit to be hand held. The
electronic components consist mainly of a
stabilized power supply for the thermal
conductivity bridge and an amplifier to
increase and measure the amount of
bridge unbalance. The electrical power
source can be either batteries or line
current. A four-step attenuator makes it
possible to vary the sensitivity of the
meter response by two decades.
Leakage Sensitivity of Hot
Wire Bridge Thermal
Conductivity Tester
The minimum detectable leak, in terms of
quantity of tracer gas per unit time,
depends on the rate of flow of the gas
through the leak detector and the
minimum concentration to which the hot
wire bridge detector will respond. By
reducing the rate of flow, smaller leaks
can be detected. However, there is a
practical limit, because it is important in
leak location that the detector should
FIGURE 35. Thermal conductivity leak detector using two hot
wire detectors in a Wheatstone bridge arrangement.
Fan
Filament
Probe tip intake
Motor
Thermal
conductivity
bridge
Reference tube
266
Leak Testing
respond quickly when the probe traverses
the position of the leak. Reducing the rate
of flow of tracer gas lengthens the
response time and beyond a certain point
the indications from the leak detector
become meaningless.
The detector shown in Fig. 35 can
detect a 60 µL·L-1 concentration of
hydrogen gas. This gives a response at one
tenth of full scale, with a pumping speed
for the probe of 0.13 cm3·s–1
(0.5 in3·min–1). The instrument will detect
an 8 × 10–7 Pa·m3·s–1 (8 × 10–6 std cm3·s–1)
hydrogen leak. With argon, which has a
much lower thermal conductivity
difference from air, only a 1.3 × 10–5
Pa·m3·s–1 (1.3 × 10–4 std cm3·s–1) leak can
be detected.
When testing with the hot wire bridge
thermal conductivity detector, the
atmosphere must be free from tracer gas.
If a system with very large leaks is being
tested, the local atmosphere may become
contaminated with tracer gas. Although
this will be inherently balanced out by
the reference circuit, ultimate leakage
sensitivity is bound to decline.
Advantages and Limitations of
Hot Wire Bridge Leak Detector
The relatively low operating temperature
of the filaments makes the hot wire bridge
leak detector quite safe to use under most
industrial conditions. The functional life
and long-term stability of the sensing
elements are good. The only effect that
has been noted after long periods of
operation under industrial conditions was
the accumulation of a dust deposit in the
intake line, which was easily removed.
Unfortunately, this versatility is also a
disadvantage. Because of a lack of
selectivity, this instrument can not be
operated at high sensitivity in atmospheres
contaminated with other gases.
The thermal conductivity bridges used
in these detectors do not actually measure
thermal conductivity. Because of their
structure, the readings obtained with
these detectors are dependent on tracer
gas thermal conductivity combined with
density, accommodation coefficient and
viscosity. Therefore, the values of
sensitivity inferred from thermal
conductivities of Table 7 are not absolute,
but merely an indication of the expected
general trend in the results.
A thermal conductivity detector,
similar to that of Fig. 35, uses a fourelement wire bridge. This bridge was also
found useful for vacuum leak detection.
The sensitivity of this type of leak detector
was improved by use of thermistors, with
their higher thermal coefficient of
resistance, instead of wire elements. These
detectors were tested in submarine service,
where they were found useful in detecting
leaks of a variety of gases.
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
PART 9. Leak Testing of Vacuum Systems by
Ionization Gage or Pump Techniques
Ionization Gage Technique
of Leak Testing of Vacuum
Systems
The ionization gage technique of leak
testing of pumped vacuum systems
involves spraying the suspected leak area
with tracer gas and observing any pressure
change indicated on an ionization gage.
Any gage that measures ionization of the
gas may be used; this can be either a hot
cathode gage, a cold cathode gage or even
an ion pump.
In ionization gages, ionization current
depends on the probability of ionizing
collisions. With all other variables held
constant, this probability of ionization
varies from one gas to another. When the
tracer gas is applied to the leak, some of
the gas in the gage is replaced by tracer
gas that causes an ionization current
either lower or higher than the steady
ionization current due to the prevailing
pressure in the system in the absence of
tracer gases.
As long as the leaks being located are
the ones that limit the system pressure,
the ionization gage technique may be
applicable to very low pressures and/or
very low leakage rates. It has been used
for location of leaks in ultrahigh-vacuum
systems. On very small volume systems,
this technique is reported to be more
sensitive than the mass spectrometer leak
detector.
Use of Ionization Gages As
Leak Detectors for Vacuum
Systems
As described above, ionization gages
respond differently to different gases. For
example, if first air and secondly helium
are admitted through a small (molecular
flow) leak into a system using diffusion
pumps, then the ionization gage response
to the helium will be about 15 to
20 percent of the response to air. In this
case the actual pressure in the system will
be virtually unchanged. This follows
because both the leakage rate and the
pumping speed vary in the same way.
Both are inversely proportional to the
square root of molecular mass. The
decreased response for helium is due to
the fact that the ionization potential of
helium is much higher than the
ionization potentials of nitrogen, oxygen
or air. On the other hand, the ionization
potentials of hydrogen and carbon are
somewhat lower than that of air and
indeed the response of an ion gage to
hydrogen and hydrocarbon compounds
such as acetone, alcohol or butane is
greater than that of air. The application of
this behavior to leak detection is obvious.
In practice, one usually adjusts the gridcurrent control until the ion gage reads
near full scale to obtain maximum
sensitivity. Then the system is probed
with one of the tracer gages or vapors
mentioned while monitoring the reading
of the ionization gage.
Effect of Tracer Gas
Properties on Ionization
Leak Test Sensitivity
It is desirable that the ionization
efficiency of the tracer gas be as different
as possible from that of the background
gas (air). In general, gage sensitivity
increases with the number of electrons in
the molecule. Examination of ion gage
sensitivities suggests that the best gases
for this technique are either the low
molecular weight gases such as hydrogen,
helium and neon or the high molecular
weight vapors such as acetone, ether and
alcohol. In using the vapors, care must be
taken that they do not plug the leak. In
some cases, response may be delayed
because of adsorption of vapors on the
interior surface of the leak.
Care must be taken that the tracer gas
does not permanently react and change
the gage sensitivity. For example, applying
carbon dioxide for a time can change the
sensitivity of a Penning gage. The
discharge current decreases about 30 to
40 percent probably because of a film of
carbonates on the electrodes. This general
technique can be modified in several
ways. Instead of an ionization gage, an
ion pump may be used. Selectivity of the
gage to the tracer gas may be increased by
use of a double gage setup, where a gage is
positioned so that it is selective only to
the tracer gas. Another modification of
this technique is to use the poisoning
effect of oxygen on the emission of
electrons from a tungsten filament.
Leak Testing of Vacuum Systems
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
267
Sensitivity of Ion Gage
Leak Detection in Vacuum
Systems
Several conditions can reduce the
sensitivity of the ion vacuum gage leak
detection technique. If several leaks are
present in the system, the differential
response of the gage will be smaller than
for a single leak. The response time of a
large systems may be comparable with the
fluctuations or drift that may be present
on the normal gage reading. In such cases
it is difficult to tell when a leak has
actually been encountered. If a leak is
definitely suspected in one location, the
signal-to-noise ratio can be improved
somewhat by alternately probing with
helium and acetone.
The sensitivity of the ionization gage
technique can be greatly improved by
commercially available leak detection
devices that attach to the recorder
terminals of most ionization gage and ion
pump circuits.
Leak Detector with
Magnetron Ionization
Gages
Another leak detector uses two magnetron
ionization gages enclosed as a unit of the
same general dimensions as the mass
spectrometer leak detector analyzer
section. The two ionization gages are
connected in series, with the second gage
cryogenically trapped. The two gages are
balanced on a bridge circuit. Tracer gas
changes the current of the first gage, but
is condensed and therefore does not affect
the second gage. With two gages,
background pressure variations do not
affect the detector. The leakage sensitivity
of this magnetron ionization detector is
reported to be 10–11 Pa·m3·s–1
(10–10 std cm3·s–1).
sensitive microammeter, either of which is
provided with a suitable shunting circuit.
In a stable vacuum, constant current flows
through the gage tube and the
potentiometer, creating a steady voltage
drop across the potentiometer. The
battery provides a reference voltage and
the potentiometer can be adjusted to give
a null indication on the galvanometer.
The shunting switch is left closed until
this adjustment is made.
As shown in Fig. 36, the null set
potentiometer devices can compensate
the gage current due to the air leak (i.e.,
provide a counter current adjusted to give
a null reading) and then amplify any
variations from null. The result is a great
magnification of pressure variations too
small to be detected on the meter of the
ion gage. Noise and drift variations,
which are magnified as well, set the
practical limit to the sensitivity obtained
by using these devices. Small leaks can
sometimes be located in the presence of a
pressure drift if the output of the ion gage
leak detector is monitored with a strip
chart recorder. The location of the leak is
indicated by the change in slope of the
drift curve. For stable systems, the ion
gage leak detector can detect a 1 percent
change in the pressure reading of the gage
circuit. Because this sensitivity approaches
or exceeds that of the helium mass
spectrometer leak detector for pressures
below 1 µPa (10 ntorr), the ionization
gage technique is often used with vacuum
systems operating in the ultrahigh
FIGURE 36. Null balance circuit for leak location with an
ionization gage leak detector.
–
Direct current
power supply
+
Ionization
gage
Differential Ionization
Gage Leak Detection
Instrumentation
To obtain adequate leakage sensitivity
with the ionization gage technique, the
background ionization current may be
nulled using a sensitive difference
amplifier or a galvanometer with backing
off voltage control, so that very small
changes in ionization current are
detected. An example of a circuit for such
testing is shown in Fig. 36. The indicating
instrument has been replaced with a
potentiometer. The null-balance
instrument can be a galvanometer or a
268
Leak Testing
Shunt
To vacuum
system
Potentiometer
Null indicator
Reference voltage
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
vacuum range. But if very small leaks
must be found at moderate vacua, about
100 µPa (1 µtorr), as for example in the
leak testing of an ultrahigh vacuum
system bakeout, then a mass spectrometer
detector must be used.
Selective Tracer Gas
Transmission Leak Testing
with Ionization Gage
The sensitivity of an ion gage to tracer gas
can be increased if air is excluded and the
tracer gas only is selectively brought to
the ion gage. If this is done, the gage will
not respond to extraneous pressure
changes. Selectivity can be increased by
use of a selective membrane or a
cryogenic trap in front of the gage. For
example, palladium metal passes only
hydrogen gas. On the other hand, silica
gel passes not only hydrogen but the
noble gases (helium, neon and argon).
Neither palladium nor silica gel will pass
air through the barrier wall. A cryogenic
cold trap can collect hydrocarbon vapors
that condense with it, so they cannot
form interfering carbon layers on barriers
of ionization gage components.
Palladium Barrier
Ionization Gage for
Detecting Leaks in Vacuum
Systems
The palladium barrier gage is typical of
several that have the property of selective
allowing hydrogen to pass into a vacuum
gage, to the exclusion of all other gases. It
uses the fact that hot (about 800 °C or
1470 °F) palladium metal is permeable to
hydrogen but not to other gases. As
shown in Fig. 37, the palladium barrier
gage is in essence an ionization gage with
a palladium barrier between it and the
vacuum system. The palladium is heated
either by electron bombardment or by
conduction from a hot filament. The gage
is evacuated, sealed off and gettered to
achieve a very low pressure in the gage
itself. The gage can be placed in the
foreline of the system; because only the
hydrogen passes through the barrier, the
pressure in the gage is just the partial
pressure of this hydrogen tracer gas alone.
It is claimed that this device can detect
changes as small as 3 µPa (20 ntorr) in the
partial pressure of hydrogen and some
claim to have detected leaks as small as
5 × 10–11 Pa·m3·s–1 (5 × 10–10 std cm3·s–1).
However, sensitivities corresponding to
leakage rates in the range 10–7 to
10–8 Pa·m3·s–1 (10–6 to 10–7 std cm3·s–1) are
more normal in actual practice.
Leakage Sensitivity of Palladium
Barrier Ionization Gage
The direction and rate at which hydrogen
passes through the palladium barrier
depends on the hydrogen pressure
differential across the barrier. Thus,
hydrogen in the gage volume can be
removed by reducing the external
hydrogen pressure below the internal
value. The gage can detect a pressure
change of about 3 µPa (20 ntorr), but
must be operated under carefully
controlled conditions to achieve this
sensitivity. Use has been made of a
hydrogen generator consisting of a hot
tungsten filament that decomposes oil
vapors present in the vacuum system. To
obtain maximum leak detection
sensitivity, it is sometimes found
necessary to maintain a hydrogen partial
pressure in the system of about 40 µPa
(0.3 µtorr) by glowing the tungsten
filament at temperature of about 800 °C
(1470 °F).
Precautions with Palladium Barrier
Ionization Gage
It is necessary to place a liquid nitrogen
trap between the palladium barrier
ionization gage leak detector and the rest
of the system to exclude hydrocarbons
and water vapor from the gage. These
vapors dissociate at the hot palladium
surface to give hydrogen, which produces
a spurious response. In addition, the
cracked hydrocarbons build up a carbide
layer on the palladium, which reduces its
permeability. It is also desirable to use a
turbomolecular pump with oil free
bearings rather than an oil diffusion
pump in the vacuum system; otherwise,
the hydrogen that results from the
decomposition of diffusion pump oil gives
rise to an unstable background ion
current in the gage. In a system
containing multiple leaks, oxygen in the
air entering the undetected leaks
FIGURE 37. Palladium barrier ionization gage.
Cylindrical
ion
collector
Glass
envelope
Tube
Heater
Cathode
Palladium
anode
Earth wire
Leak Testing of Vacuum Systems
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
269
combines at the hot palladium surface
with the hydrogen entering through a
leak that is being probed. If there is an
excess of oxygen, all hydrogen will react
with the oxygen before it can pass
through the barrier and will therefore be
undetected. Under these circumstances, a
controlled leak of hydrogen could be
admitted to the system to take up the
oxygen.
If air is admitted to the ion gage, the
palladium becomes oxidized even if it is
cold. Whenever this occurs, 2 to 3 h of
run-in time is required to obtain
reproducible results on duplicate runs.
Therefore, even if the gage is not in use,
the forepumps should be operated
continuously to prevent air contact with
the palladium. If the gas is left exposed to
the atmosphere, several warm-up runs
should be made to allow hydrogen to pass
through the calibrated leaks and be
pumped down between successive runs.
Vacuum Leak Testing with
Cryogenically Trapped
Gage with Silica Gel
Absorbent instead of
Palladium
It is possible to use an absorbent to pass
the tracer gas and block air. Silica gel,
outgassed at 300 °C (570 °F) and then
cooled to liquid nitrogen temperatures, is
commonly used for this purpose. Under
these circumstances, silica gel readily
passes hydrogen and the noble gases
(helium, neon, argon), but not air. The
system uses a cold cathode gage and
hydrogen. The gage is separated from the
system by a liquid nitrogen cold trap
filled with silica gel.
Sensitivity of Silica Gel Absorbent
Leak Testing
When silica gel is used in the cold trap,
the ionization gage leakage sensitivity is
claimed to be about a hundred times
greater than that of the palladium
hydrogen system. However, several hours
are required to measure leakage rates of
the order of 10–13 Pa·m3·s–1
(10–12 std cm3·s–1). Careful degassing of
the leak detector and the tube to be tested
is necessary. One advantage claimed for
silica gel is a long usage time before it has
to be degassed again. The increased
sensitivity of silica gel is claimed to be
due to less gas evolution from the gel
than from heated palladium, which
results in lower pressures. This detector,
although very sensitive, is limited by long
pump down times.
270
Leak Testing
Ion Pump Technique of
Vacuum System Leak
Detection
Cold cathode, gas discharge ion pumps
are convenient instruments for leak
location. An ion pump acts not only as a
pump but also as an effective pressure
gage, because the pump current is
proportional to the number of molecules
being pumped. The pump current is also
dependent on the ionization efficiency of
the gas molecules being pumped. The
pumping speed is dependent on the
ionization efficiency of the gas molecules
being pumped. The pumping speed is
dependent on the molecular chemical
reactivity rather than the molecular
weight, so the response of an ion pump to
a tracer gas will be different from the
response of an ionization gage. A typical
arrangement for ion pump leak testing of
evacuated systems is shown in Fig. 38.
Effect of Tracer Gas on Leakage
Response of Ion Pump
The response of an ion pump to various
probe gases is shown in Fig. 39. As may be
seen from those curves, the response
differs with time, not only in magnitude,
but also occasionally in sign. The best
gases for leak location using an ion pump
seem to be argon, oxygen and carbon
dioxide.
The pumping speed of an ion pump
depends strongly on the chemical activity
of the gas being pumped. Unlike a
turbomolecular pump or a diffusion
pump, the pumping speed of an ion
pump varies with chemical species rather
than with molecular mass. The actual
pressure in an ion pumped vacuum
system will thus vary as different gases are
introduced via a molecular flow leak.
FIGURE 38. Ion pump leak detector arrangement.
Tracer probe
Ion pump
gage circuit
Leak
System
being
tested
Thermocouple
gage
P
Ion pump
V1
V2
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
Therefore, if the ion gage is mounted on
an ion pumped system, the change of its
indicated pressure in response to a change
in gas composition will be markedly
different from that of the same gage on a
diffusion pumped system. It is customary
to use the ionization current in an ion
pump as a measure of the pressure in the
pump. The response of such an ion pump
pressure gage is similar to that of an
external thermionic ionization gage. Both
types of gage show an increase in pressure
when either argon or helium enters the
vacuum system. A pressure decrease is
indicated when oxygen or carbon dioxide
enter. Thus, these gases can be used to
detect leaks in ion pumped systems in the
same manner as the ionization gage
described just previously. However, note
that the two types of pumps give opposite
responses for helium.
The sensitivity of an ion gage leak
detector on a system using ion pumping
is shown in Fig. 41 as a function of
pressure.
FIGURE 40. Schematic circuit diagram of an ion pump leak
detector.
Recorder
output
Ion pump
gage circuit
Null Set Circuit for Ion Pump Leak
Detector
FIGURE 39. Response of an ion pump gage indication to leaks
of various gases.
0.6
Argon
0.5
Gage response (relative units)
0.4
Helium
0.3
0.2
Hydrogen
0.1
0.0
Leakage rate
indicator
Null set
potentiometer
FIGURE 41. Minimum detectable leakage rate as a function of
pressure for vacuum systems with ion pumping.
10–5 (10–4)
10–6 (10–5)
Mass flow rate, Pa·m3·s –1 (std cm3·s –1)
The response amplifier type of ion gage
leak detector circuit sketched in Fig. 40
can be used with the recorder output of
the circuit associated with either an ion
pump or a thermionic ion gage. The
pressure fluctuations (noise) or an ion
pumped system are usually somewhat less
than for a turbomolecular or diffusion
pumped system, unless the ion pump is
experiencing argon instability (burping).
Stable
direct
current
amplifier
10–7 (10–6)
10–8 (10–7)
10–9 (10–8)
10–10 (10–9)
10–11 (10–10)
Hydrogen with added pumping
10–12 (10–11)
Helium with added pumping
10–13 (10–12)
10–9
– 0.1
– 0.2
– 0.3
(10–13)
– 0.4
10–8
10–7
10–6
10–5
10–4
10–3
10–2
(10–12)
(10–11)
(10–10)
(10–9)
(10–8)
(10–7)
(10–6)
Pressure, Pa (lb f·in.–2 × 1.45)
Oxygen or carbon dioxide
– 0.5
Legend
– 0.6
0
1
2
3
4
5
6
7
Time (relative units)
8
9
10
=
=
=
=
=
400 L·s–1 (850 ft3·min–1)
125 L·s–1 (265 ft3·min–1)
75 L·s–1 (160 ft3·min–1)
40 L·s–1 (85 ft3·min–1)
8 L·s–1 (17 ft3·min–1)
Leak Testing of Vacuum Systems
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
271
Leakage Sensitivity of Ion Pump
Technique
Leaks in the 10–12 Pa·m3·s–1 range may be
located with an ion pump. This is a
conservative estimate of the sensitivity;
the current changes being measured are
several orders of magnitude greater than
the corresponding mass spectrometer ion
currents. With the ion pump leak detector
system shown in Fig. 38, the procedure is
to evacuate an ion pump and keep it
operating at low pressure with the valve
V1 closed. The system to be leak tested is
first evacuated by a mechanical pump to a
pressure of about 1 Pa (7 mtorr). Valve V1
is then opened and V2 closed until an
equilibrium pressure is reached (a few
minutes). When the leak is probed with
argon, the ion pump current should
increase rapidly, presumably due to the
low speed of the pump for argon. Probing
with hydrogen and oxygen causes a
reduction in pressure, because these gases
are pumped more rapidly than air. With
helium used as the search gas, the
sensitivity is lower than for argon.
Leaks as small as 10–11 Pa·m3·s–1 (or
–10
10 std cm3·s–1) are located using the ion
pump technique. Leaks between 10–4 and
10–6 Pa·m3·s–1 (10–3 and 10–5 std cm3·s–1)
could be located by partial opening V1
and by having V2 opened sufficiently to
avoid a pressure increase in the system
during the leak testing procedure. Leaks of
10–6 to 10–7 Pa·m3·s–1 (10–5 to 10–6 std
cm3·s–1) could be determined a few
minutes after opening V1 and closing V2.
Leaks smaller than 10–9 Pa·m3·s–1
(10–8 std cm3·s–1) required a longer time,
depending on the volume and outgassing
properties of the item under test.
rise to a partial air pressure of 10 µPa
(0.1 µtorr) is readily detected when
probed with oxygen. The detection circuit
used is a modified ionization gage control
unit. The filament is heated by a regulated
power supply, but is not emission
regulated. For stable operation of this type
of detector, using thoria coated tungsten
filaments, it is best to reduce the thoria to
thorium at the beginning of the test by
heating the filament for a few seconds to
a temperature of 2400 K (3860 °F).
Sensitivity Characteristics of
Thermionic Electron Emission
Oxygen Leak Detector
The greatest sensitivity to oxygen tracer
leakage is at an operating temperature just
below 1900 K (2960 °F), when the
tungsten surface is partly covered with
thorium. This can be obtained only when
leaks of 10–10 Pa·m3·s–1 (10–9 std cm3·s–1)
or less are remaining in a well baked
system pumped at a speed of 10 L·s–1
(21 ft3·min–1). The filament can become
desensitized when it becomes carburized.
It is because of the danger of carburization
in the presence of hydrocarbon vapors
and because of the influence of residual
water vapor on the emission of electrons
from the thoriated tungsten, that the
detector is not very suitable for use in leak
testing of unbaked vacuum apparatus. If a
filament becomes carburized accidentally
it must be replaced; no thermal treatment
cycle will bring it to a sensitive state
again. But in a well baked system,
thoriated tungsten filaments can, if
necessary, always be restored to a desired
state of sensitivity again by a short period
of running at a temperature of about
2400 K (3860 °F).
Leak Detection by
Reduction of Thermionic
Electron Emission by
Oxygen Tracer Gas
A very sensitive means of locating leaks in
vacuum systems is to observe the
temperature limited emission of electrons
from a heated tungsten filament in a
vacuum. When a stream of oxygen tracer
gas is blown over the outside of a leak,
the resulting increase in oxygen pressure
within the vacuum system causes the
filament’s emission to drop. Although the
principle has been known for a long time
and various circuits have been developed
for its use, this technique has not been
extensively used. An instrument in which
the grid of a triode ionization gage is
connected externally to the collector to
form a diode is used to detect oxygen
admitted to the apparatus under
controlled conditions. A leak that gives
272
Leak Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
References
1. Marr, J.W. Leakage Testing Handbook.
Report No. CR-952. College Park, MD:
National Aeronautics and Space
Administration, Scientific and
Technical Information Facility (1968).
2. Leybold Inficon Incorporated. Product
and Vacuum Technology Reference Book
[1995/96]. East Syracuse, NY: Leybold
Vacuum Products Incorporated and
Leybold Inficon Incorporated (1995).
Leak Testing of Vacuum Systems
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
273
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
C
7
H A P T E R
Bubble Testing
Gerald L. Anderson, American Gas and Chemical
Company Limited, Northvale, New Jersey
Charles N. Jackson, Richland, Washington
Robert W. Loveless, Nutley, New Jersey
Charles N. Sherlock, Willis, Texas
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
PART 1. Introduction to Bubble Emission
Techniques of Leak Testing
Principles of Bubble
Testing for Leaks
In leak testing by the bubble test
technique, a gas pressure differential is
first established across a pressure
boundary to be tested. A test liquid is
then placed in contact with the lower
pressure side of the pressure boundary.
(This sequence prevents the entry and
clogging of leaks by the test liquid.) Gas
leakage through the pressure boundary
can then be detected by observation of
bubbles formed in the detection liquid at
the exit points of leakage through the
pressure boundary. This technique
provides immediate indications of the
existence and location of large leaks,
10–3 to 10–5 Pa·m3·s–1 (10–2 to 10–4 std
cm3·s–1). Longer inspection time periods
may be needed for detection of small
leaks, 10–5 to 10–6 Pa·m3·s–1
(10–4 to 10–5 std cm3·s–1), whose bubble
indications form slowly.
In bubble tests, the probing medium is
the gas that flows through the leak due to
the pressure differential. The test
indication is the formation of visible
bubbles in the detection liquid at the exit
point of the leak. Rate of bubble
formation, size of bubbles formed and rate
of growth in size of individual bubbles
provide means for estimating the size of
leaks (the rate of gas flow through leaks).
Classification of Bubble
Test Techniques According
to Test Liquids
Bubble test techniques for detecting or
locating leaks can be divided into three
major classifications related to the
technique of using the test liquid:
1. In the liquid immersion technique,
the pressurized test object or system is
submerged in the test liquid. Bubbles
are then formed at the exit point of
gas leakage and tend to rise toward the
surface of the immersion bath.
2. In the liquid film application
technique, a thin layer of test liquid is
flowed over the low pressure surface of
the test object. An example of this
solution film leak test is the well
known soap bubble technique used by
276
Leak Testing
plumbers to detect gas leaks. Films of
detection liquid can be readily applied
to many components and structures
that cannot be conveniently immersed
in a detection liquid. For detection of
small leaks, this liquid should form a
thin, continuous, wetted film covering
all areas to be examined.
3. The foam application technique is
used for detection of large leaks in
which the applied liquid forms thick
suds or foam. When large leaks are
encountered, the rapid escape of gas
blows a hole through the foam
blanket, revealing the leak location.
Classification of Bubble Test by
Pressure Control
Subclassifications of these basic
techniques of bubble testing refer to
different techniques for controlling the
pressure differential acting across the
pressure boundary. Several techniques are
used to raise the pressure differential and
so to increase the rate of gas leakage and
the rate of formation of bubbles.
1. Pressurize the interior volume of the
test object or system before and during
the leak test. Internal gas pressure
should be applied across the pressure
boundary before test liquid contacts
the external surface. This tends to
prevent entry of liquid into leaks,
which might possibly clog the leaks to
gas flow. Protection against hazards of
overpressure must be provided.
2. Control the heating of sealed test
objects and small components to
cause internal gas expansion. This
increases the pressure differential and
causes outward gas flow through
possible leaks in the pressure
boundary.
3. Apply a partial vacuum above the
surface of the test liquid (immersion
liquid or solution film). This reduces
external pressure to the pressure
boundary. The resultant increase in
pressure differential across the system
boundary acts to cause gas flow
through any leaks that are present.
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
Advantages of Bubble
Testing
Bubble testing has the obvious advantages
of being relatively simple, rapid and
inexpensive. It is a fairly sensitive leak
detection technique and enables the
observer to locate the exit points of leaks
very accurately. (The point of exit may
not be directly opposite the entry point of
the leak, especially in welds or castings.)
Another major advantage of bubble
testing is that very large leaks can be
detected readily. Bubble test techniques
also provide very rapid responses even for
small leaks. (Some more sensitive leak
testing techniques often have responses so
slow that a leak may be missed while
probing.) With bubble tests, it is not
necessary to move a tracer probe or
detector probe from point to point. In
immersion bubble tests, the entire
pressurized component can often be
examined simultaneously for leaks on
exposed surfaces visible to the observer. In
some cases, test components may have to
be turned over to expose the underside to
view, so that leaks from this area can be
seen. All leaks are revealed independently
in immersion bubble testing. If desired,
large leaks can be first detected with rapid
bubble test techniques. These leaks can
then to sealed before refined leak testing
apparatus is used to detect smaller leaks.
The bubble testing technique lets the
observer distinguish real from virtual
leaks. (Virtual leakage is a primary
problem in leak testing of vacuum
systems but may also be encountered
when bubble testing.) In addition, during
bubble tests it is not necessary that all
connection pipes and valves be free from
leaks. However, detection of small leaks
requires operator patience and additional
test time for bubble or foam indications
to form. Care is required to ensure that all
detectable bubble indications present are
observed. Bubble testing is satisfactory for
detecting gross leakage. With inert
probing gases and test liquids, bubble
tests are fairly safe in a combustible
atmosphere. However, this depends on
selection of proper tracer gas and test
liquids. The required level of operator
training and skill is minimal, compared
with some more complex techniques of
leak testing.
Limitations of Bubble
Techniques of Leak Testing
Conditions that interfere with bubble
emission techniques of leak testing or
limit their effectiveness include the
following: (1) contamination of test
specimen surfaces; (2) improper
temperatures of test specimen surfaces;
(3) contaminated or foaming test liquids;
(4) improper viscosities of test liquids;
(5) excessive vacuum over surface of test
liquid; (6) low surface tension of test
liquids leading to clogging of leaks;
(7) prior use of cleaning liquids that clog
leaks; (8) air dissolved in test liquids or
outgassing from corroded test surfaces,
causing spurious bubble formations; and
(9) leaks with directional flow
characteristics, intermittent or very slow
leakage or porosity leaks.
Prior bubble testing or contamination
may clog leaks and lower the sensitivity of
subsequent leak testing by more sensitive
techniques.
Effects of Test Surface
Contamination, Porosity or
Temperature
Surface contamination of the test
specimen can occur with small immersed
test parts or on scaled, dirty or greasy
surfaces of large vessels or components.
Grease, rust, weld slag, oxide films or
other surface films, as well as weld
porosity open to a surface may be sources
of bubbles giving false indications of
leakage. Temporary plugging of leaks
might also occur because of some
common manufacturing techniques such
as peening or metal smearing that closes
the openings to leaks at metal surfaces.
Leak testing must be done before
painting, galvanizing, coating or plating
of surfaces, which may plug leaks
temporarily. Difficulties can also result
when tests are performed with test
specimen surface temperatures either too
high or too low for inspection procedure
requirements.
Effects of Properties and
Contamination of Bubble Test
Liquid
Contaminated test liquids or test liquids
that foam on application can cause
formation of spurious bubbles on test
specimens, which is not related to leakage
through the pressure boundary. Incorrect
viscosity of the test fluid can also affect
formation of visible streams of bubbles at
leaks. Formation of spurious bubbles
caused by air dissolved in water or other
immersion liquids hinders detection of
bubble emission from real leaks. When
bubble tests are conducted on metallic
vessels, some bubbles can evolve from
outgassing from patches of corrosion.
Effects of Excessive Vacuum over
Bubble Test Liquid
Excessive vacuum on the low pressure side
of the pressure boundary of test objects
Bubble Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
277
could occur when using the vacuum box
pressure differential technique of bubble
testing. Excessive vacuum (absolute
pressure too low over the test liquid) can
lead to boiling of the detection liquid.
When the immersion liquid is boiling,
bubbles of vapor form throughout the
solution and typically rise to the liquid
surface. These could interfere with
operator detection and observation of
bubble formation caused by leakage. The
amount of vacuum allowed in immersion
bubble testing depends on the immersion
test liquid. It should be the maximum
vacuum attainable without causing the
test liquid to boil.
Effects of Low Surface Tension of
Bubble Test Liquid
Clogging of small leaks with leakage rates
less than 10–5 Pa·m3·s–1 (10–4 std cm3·s–1)
can result from premature application of
the test liquid, either by immersion or
film solution. Most bubble testing
solutions have a low surface tension.
Detection solutions with low surface
tension promote surface wetting. This
increases the tendency of the test liquid
to enter and block very small leaks. This
tendency can be reduced, however, if the
vessel or test component is always
pressurized before covering the surface
under test with any liquid. Clogging of
existing leaks could also occur if the test
liquid used in bubble emission tests enter
the leaks after an external vacuum is
released.
Effects of Prior Surface Cleaning
of Test Objects
Prior use of cleaning liquids on test object
surfaces can also result in clogging of
leaks. Thus, all test objects must be
thoroughly dried by heat or vacuum or
both, after cleaning with liquid solutions
before leak testing with gaseous tracers.
Effects of Porosity, Intermittent
Leaks and Check Valve Leaks
Leaks with special characteristics may
react in ways such that they cannot
always be found reliably by bubble tests.
For example, porosity leaks cannot be
detected by bubble tests if the pores are
very small. Some types of leaks may pass
gas in only one direction; if this direction
is inward, bubble tests of outside surfaces
will not detect them. With intermittent or
very slow leaks, close operator
surveillance of the test surface is often
necessary to detect bubbles.
278
Leak Testing
Importance of Cleaning Test
Surfaces after Bubble Testing
Cleaning of test object surfaces and
drying of test objects to remove all bubble
test liquids from within leaks is essential
when these same test objects are
subsequently subjected to more sensitive
leak tests with gas tracers (such as halogen
vapor or helium leak tests). The later gas
tracer leak tests could be invalidated if
prior bubble testing had clogged the leaks
with water or other liquids.
Factors Influencing the
Sensitivity of Bubble
Testing
As noted earlier in this chapter, the basic
principle of the bubble test consists of
creating a pressure differential across a
leak and observing bubbles formed in a
liquid medium located on the low
pressure side of the leak or pressure
boundary. The sensitivity of the bubble
test technique can be influenced by
factors such as (1) pressure differential
acting across the leak; (2) viscosity of
pressurizing tracer gas; (3) test liquid used
for bubble formation; (4) contamination
on surfaces being tested (i.e., paint, dirt,
oil etc. on inside or outside surface of
object being tested); (5) ambient weather
conditions (such as rain, temperature,
humidity or wind); (6) lighting in test
area; (7) test equipment; and (8) test
personnel technique and attitude.
Properties Affecting Leak Detector
Solution Performance
1. Surface tension affects the speed and
size of bubble formation. Lower
surface tension solutions form many
small bubbles and the reforming of
new bubbles. Higher surface tension
solutions slowly form very large
bubbles that are slower to break, but
usually do not reform new bubbles.
Water softener is used to reduce
surface tension.
2. Good wetting action and a large contact
angle are the result of lower surface
tension. Poor wetting action and a
small contact angle are the result of
higher surface tension.
3. Viscosity affects the size of bubble
growth. Lower viscosity solutions
produce smaller bubbles. Higher
viscosity solutions produce larger
bubbles. Glycerine may be used to
control viscosity.
4. Evaporation rate controls the amount
of test area that may be covered with
leak detector solution before the final
inspection. It is desirable therefore to
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
use a solution that has a slow
evaporation rate to be able to cover a
larger test area. Evaporation rate is also
temperature dependent with an
increase in temperature causing an
increase in evaporation rate and vice
versa.
Techniques for Attaining
Required Bubble Test
Sensitivities
As long as the pressure differential can be
maintained, the bubble test technique can
be used. However, the sensitivity of a leak
testing procedure must be adequate to
permit detection of all leaks of a certain
size and larger so that all detected leaks
can be repaired. The hole or crack that
constitutes the physical leak is usually
characterized for size of leak by the
amount of gas passing through it as
leakage. The sensitivity of a bubble test
can be increased by (1) increasing the
time allowed for bubble formation and
observation, (2) improving conditions for
observing bubble emission and
(3) increasing the amount of gas passing
through the leak.
Improving Bubble Test Sensitivity
by Better Observational
Capabilities
The actual sensitivity of a specific leak test
procedure can be improved by an increase
in observational ability. An increase in
observational ability could be attained by
the following means.
1. Position test surfaces optimally for
visual inspection.
2. Improve lighting to highlight bubble
emission clearly and use clean
translucent immersion liquids.
3. Increase time for bubble formation
and observation by test operators.
4. Eliminate false bubble indications
(caused by boiling, entrained air or
contamination of inspection liquids,
for example).
5. Decrease surface tension of the
detection liquid that causes more and
smaller bubbles to appear.
6. Reduce pressure above the inspection
liquid, which makes the individual
bubbles larger.
7. Select test site and time to provide
optimum ambient conditions, such as
temperature, wind and lighting
conditions.
8. Use leak detector solutions that are
fluorescent and colored for increased
contrast with different test surfaces.
Factors affecting operator comfort and
ability to see bubble indications must also
be considered. Tests might be postponed
until proper test conditions can be
attained.
Each of these aids to sensitivity enables
the test operator to detect the bubble
emissions from smaller leaks or to
separate the indications for closely
adjacent leaks more readily and so
improve the reliability of leak detection.
Increasing Bubble Test Sensitivity
by Raising Tracer Gas Flow Rate
Increase in sensitivity resulting from
improvements in leak test procedures are
typically attained by raising the rate of
flow of tracer gas through the existing
leaks. The increased amount of gas flow
through the leak passageway may be
attained by a change in the properties of
the gas (lower gas viscosity or lower
mass). Alternatively, the quantity of gas
passing through the leak could be
increased by applying a higher pressure
differential across the leak. This higher
differential pressure could be achieved by
a higher level of internal gas
pressurization of the vessel or component
under test, by heating the gas within a
sealed component to increase its pressure
or by reduction of the pressure acting
through the test liquid on the low
pressure side of the pressure boundary.
These techniques increase the sensitivity
of the test procedure to which the
components are subjected. They may also
result in more easily observed bubble
indications that improve the reliability
and speed of bubble testing.
Sensitivities Attainable
with Liquid Film Bubble
Testing
The actual sensitivity attained in bubble
testing depends on the control and
selection of leak test conditions that
influence factors affecting sensitivity.
Sensitivity also depends on the selection
of the test technique. The liquid
application technique (solution film
technique), in which a thin film of liquid
is applied and bubbles form in air (like
soap bubbles floating on water), is
typically used only for leak detection and
location. A leak is a physical hole; the gas
passing through it is leakage. Service
requirements or specifications for testing
may require that any detectable leakage
be taken as cause for rejection or for
repair of leaks. In this case, it is not
necessary to measure actual leakage rates
to determine the disposition of the test
items. The sensitivity of the liquid
application technique of bubble testing is
adequate for locating leaks with leakage
Bubble Testing
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
279
rates in excess of 10–5 Pa·m3·s–1 (10–4 std
cm3·s–1). The solution film procedure is
widely used on large pressurized systems
that cannot be immersed in detection
liquid. The technique is ideal for quick
detection of large to moderate size leaks
(10–2 to 10–4 Pa·m3·s–1 or 10–1 to 10–3
std cm3·s–1) at very low costs.
Sensitivities Attainable
with Immersion Bubble
Testing
In bubble testing by the immersion
technique, test sensitivity depends on
operating conditions and selection of
both the tracer gas and the test liquids.
Other factors can also change the test
sensitivity actually attained. With certain
combinations of tracer gases and
detection liquids, sensitivities of 10–8
Pa·m3·s–1 (10–7 std cm3·s–1) have been
attained with calibrated leaks operating
under laboratory conditions. Under
excellent industrial immersion bubble
testing conditions, maximum sensitivity
of bubble testing is in the range of 10–5 to
10–6 Pa·m3·s–1 (10–4 to 10–5 std cm3·s–1).
Operator Training and Motivation
to Maintain Bubble Test
Sensitivity
The sensitivity of bubble testing is hard to
define because it also depends on the
observation and alertness of the leak test
operator. Practically, under excellent
industrial test conditions, there is no
question that leakage of 10–6 Pa·m3·s–1
(10–5 std cm3·s–1) can be observed by the
immersion bubble testing procedure.
However, it is a different matter when
operators do not know that a leak exists
and have to examine a long weld seam for
a possible bubble. Conceivably, they
might not wait long enough for the
bubbles to form or they might fail to look
carefully after sufficient time at every
portion of every area where a potential
leak might exist. Thus, optimum bubble
observation conditions and continuing
training and motivation of bubble test
operators to achieve and maintain their
best observational capabilities are essential
if the reliability and sensitivity of bubble
testing are to be ensured.
Effects of Test Pressures on Bubble
Formation
Because a minimum pressure is required
to form a bubble in a liquid, bubble
testing sensitivity depends on the pressure
differential acting across a leak. Bubble
testing sensitivity increases with an
increase of pressure across a leak.
280
Leak Testing
Sometimes, it is possible for the operator
to estimate that a certain rate of leakage is
observed because a bubble of a particular
volume is being observed. However, this
type of leakage rate estimation can be
inaccurate on very small leaks because of
the finite solubility of the tracer gas in the
bubble test liquid. It is theoretically
possible for a small leak to exist where the
tracer gas from a capillary leak dissolves in
the test liquid so fast that no leakage
bubble indication is visible. Special
techniques that serve to increase the
pressure differential across the leaks can
be used to increase bubble testing
sensitivity. Sensitivity improvements
resulting from such special techniques are
described in the discussions of each
individual technique in this chapter.
Preparation of Test Objects
for Bubble Testing
Before bubble testing, test objects must be
prepared to ensure that surface
contamination, liquid blockage of leaks,
protective coatings, sources of gas
emission, uncovered openings and other
conditions that could interfere with
effective leak testing have been properly
corrected or controlled. In addition, safety
precautions are required when
pressurizing vessels, components and
systems for leak testing. Otherwise,
excessive pressure may destroy the test
object or injure the test operator. Typical
requirements for precision leak tests in
aerospace and general industry
specifications may serve as illustrative
examples of factors to be considered in
various applications.
Precleaning of Test Object
Surfaces before Bubble Testing
Before leak testing by bubble techniques,
the test object surface areas to be tested
must be free of oil, dirt, grease, paint and
other contaminants that might mask a
leak. Surface contamination of the test
item in the form of grease, loose paint,
rust, weld slag or chemicals may become a
source of bubbles, giving false indications
of a leak. Temporary plugging of leaks
might also occur because of common
manufacturing techniques. Leak testing
must be done before painting or plating
of test objects or else such coverings must
be removed to expose leak openings and
ensure absence of leak blockage. Tests
must not be performed on grease filled
components. Any test object condition
that could lead to contamination of the
bubble test detection fluid or that could
cause foaming of the inspection liquid
should not be permitted. Foaming creates
Copyright by ASNT (all rights reserved). Licensed to Blaine Campbell, 291463, 9/17/2016 3:29:48 PM EST. Single User License only. Copying, reselling and
networking prohibited.
spurious surface bubbles on the test
specimen.
Whenever feasible, bubble tests should
be performed before any other tests where
gas is the pressurizing medium. Any
possible clogging of leaks by prior
exposure to liquids (as by prior
hydrostatic pressure tests, surface cleaning
with liquid agents or storage in damp
environments subject to condensation of
water vapor) must be avoided. When test
surfaces have been previously exposed to
liquids such as hydrostatic tested castings,
this surface condition must be corrected
by careful drying (with heat or vacuum or
both) to remove liquid that may be
clogging the leaks.
In addition, castings to be coated after
hydrostatic testing with synthetic rubber
or rubbery coatings that require
vulcanizing after application with heat
must be dried carefully to remove any
moisture that may have penetrated into
porosity or other casting defects. Failure
to remove from these openings the water
that did not leak on hydrostatic testing
will cause the coating to blister and fail
when moisture in cavities tries to escape
during the vulcanizing of the coating.
Sealing of Openings in Vessels
and Test Objects before Leak
Testing
Leak tests must often be performed on
vessels, pipe sections, valves and other
components or system elements that have
intentional openings such as at flanges,
threaded holes, instrument connections
and points of attachment to other
elements of fluid containment systems.
All such openings must be sealed using
plugs, covers, sealing wax, pipe caps or
other components or materials that can
be readily and completely removed
following completion of leak testing.
Except when using back pressurizing
techniques, a gas inlet should be provided
by attaching a valve to one of the test
covers on all items pressurized or
subjected to vacuum during leak testing.
For the back pressuring techniques, a
calibrated pressure gage and valve should
be provided on the pressurizing chamber.
Check of Test Object and
Equipment before Applying
Pressure or Vacuum
The